PPMI_Baseline_Evaluation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "PPMI_Baseline_Evaluation.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "Hc3_9QpSs339"
      ],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python36564bitbaseconda0c46a7982bbc4f80bd0644b64cf7593c",
      "display_name": "Python 3.6.5 64-bit ('base': conda)"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MtHYrhWlClsR"
      },
      "source": [
        "#**Baseline Evaluation for Early Diagnosis of Parkinson’s Disease Progression**\n",
        "> **Author:** Dharesh Vadalia\n",
        "\n",
        "\n",
        "---\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TyxgWz9jrMHR"
      },
      "source": [
        "##**Introduction**\n",
        "Parkinson's Disease is a neuro degerative disorder  caused due to loss of dopaminergic neuron in the substantia nigra located at midbrain, disturbing the patients control over motor movements and non-motor senses. During the early stages of Parkinson's Disease, patients may experience bradykinesia, resting tremors, anxiety, disturbance in sleeps, depression, vocal impairment or fatigue.These symptoms worsen gradually, causing permanent damage to patient’s quality of life. Patient at an adult age of 60 and above tends to develop visible disabilities, such as postural instability, freeze of gait and muscle rigidity.\n",
        "\n",
        "\\\n",
        "**Objective** \\\n",
        "1.  Study various clinical assessment techniques designed by makers of PD biomarkers and identify impact factor of various contributing clinical cohorts for building machine learning model.\n",
        "2.  Predict risk of PD progression based on evaluation of baseline charachteristic captured for patients. \n",
        "\n",
        "\\\n",
        "**Research Question**\\\n",
        "How well can combined analysis of clinical biomarkers help identify risk of Parkinson’s Disease progression at prodromal stage?\n",
        "\n",
        "\\\n",
        "**Source of Data**\\\n",
        "Parkinson's Progression Makers Initiative (PPMI) (http://www.ppmi-info.org), Michael J. Fox Foundation for Parkinson’s Research.\\\n",
        "Data is fetched from PPMI data repository via dedicated pypmi API package (https://pypmi.readthedocs.io/en/latest/index.html)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6UnDkKTrr22L"
      },
      "source": [
        "##**Implementation**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Hc3_9QpSs339"
      },
      "source": [
        "###Importing Necessary Libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D9m6mhBsY67b",
        "outputId": "41e2d706-6ca4-41b5-be6b-5d375c3d6429",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 332
        }
      },
      "source": [
        "!pip install pypmi"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting pypmi\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/e6/c6/4dd75b7411571f67a61c177c19eb3c2136a938a827d1e4ab1875484eaca8/pypmi-0.3-py3-none-any.whl (42kB)\n",
            "\u001b[K     |████████████████████████████████| 51kB 1.7MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy>=0.15 in /usr/local/lib/python3.6/dist-packages (from pypmi) (1.18.5)\n",
            "Requirement already satisfied: pandas>=0.21 in /usr/local/lib/python3.6/dist-packages (from pypmi) (1.0.5)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from pypmi) (4.41.1)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.6/dist-packages (from pypmi) (1.4.1)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from pypmi) (2.23.0)\n",
            "Requirement already satisfied: python-dateutil>=2.6.1 in /usr/local/lib/python3.6/dist-packages (from pandas>=0.21->pypmi) (2.8.1)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas>=0.21->pypmi) (2018.9)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->pypmi) (2020.6.20)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->pypmi) (3.0.4)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->pypmi) (1.24.3)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->pypmi) (2.10)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.6.1->pandas>=0.21->pypmi) (1.15.0)\n",
            "Installing collected packages: pypmi\n",
            "Successfully installed pypmi-0.3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5Fxum13kc47K",
        "outputId": "12649d25-fd12-4e19-acf4-91bc57693baa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        }
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import pypmi\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "from sklearn.decomposition import PCA as sklearnPCA\n",
        "from sklearn import decomposition\n",
        "from sklearn import mixture\n",
        "import itertools\n",
        "from sklearn import metrics\n",
        "%matplotlib inline\n",
        "sns.set()\n",
        "import random\n",
        "random.seed(123)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z4fyeYQGtSE-"
      },
      "source": [
        "###Data Loading and Pre-Processing\n",
        "\n",
        "Downloading study data from PPMI repository via pypmi ***fetch_studydata()*** API.\\\n",
        "\\\n",
        "**Note: In case of poor netwrok connection API connection may timeout while downloading the dataset. In such case re-run the line of code. Download file size is approx 20 MB"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8mb_PEJDZz8J",
        "outputId": "214b3aa7-9ffe-4ef7-9d17-c7268ba57a54",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "pypmi.fetch_studydata('all', user='{your ppmi user id}', password='{your password}')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fetching authentication key for data download...\n",
            "Requesting 113 datasets for download...\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Fetching data file(s): 20.3MB [00:37, 535kB/s]\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/AV-133_Image_Metadata.csv',\n",
              " '/content/AV-133_Imaging.csv',\n",
              " '/content/AV-133_SBR_Results.csv',\n",
              " '/content/AV-133_Subject_Eligibility.csv',\n",
              " '/content/Adverse_Event_Log.csv',\n",
              " '/content/Benton_Judgment_of_Line_Orientation.csv',\n",
              " '/content/Blood_Chemistry___Hematology.csv',\n",
              " '/content/Center-Subject_List.csv',\n",
              " '/content/Cognitive_Categorization.csv',\n",
              " '/content/Clinical_Diagnosis_and_Management.csv',\n",
              " '/content/Clinical_Labs.csv',\n",
              " '/content/Code_List.csv',\n",
              " '/content/Cognitive_Assessments.csv',\n",
              " '/content/Conclusion_of_Study_Participation.csv',\n",
              " '/content/Concomitant_Medications.csv',\n",
              " '/content/Randomization_table.csv',\n",
              " '/content/Contact_Information_FOUND.csv',\n",
              " '/content/Contact_Information_Brain_Bank.csv',\n",
              " '/content/Current_Biospecimen_Analysis_Results.csv',\n",
              " '/content/Current_Medical_Conditions_Log.csv',\n",
              " '/content/DNA_Sample_Collection.csv',\n",
              " '/content/DTI_Regions_of_Interest.csv',\n",
              " '/content/DaTSCAN_SPECT_Visual_Interpretation_Assessment.csv',\n",
              " '/content/DaTscan_Imaging.csv',\n",
              " '/content/DATScan_Analysis.csv',\n",
              " '/content/Data_Dictionary.csv',\n",
              " '/content/Deprecated_Biospecimen_Analysis_Results.csv',\n",
              " '/content/Derived_Variable_Definitions_and_Score_Calculations.csv',\n",
              " '/content/Diagnostic_Features.csv',\n",
              " '/content/Epworth_Sleepiness_Scale.csv',\n",
              " '/content/FBB_Analysis_Data.csv',\n",
              " '/content/FBB_Metadata.csv',\n",
              " '/content/FOUND_RFQ_Alcohol.csv',\n",
              " '/content/FOUND_RFQ_Anti-Inflammatory_Meds.csv',\n",
              " '/content/FOUND_RFQ_Caffeine.csv',\n",
              " '/content/FOUND_RFQ_Calcium_Channel_Blockers.csv',\n",
              " '/content/FOUND_RFQ_Female_Reproductive_Health.csv',\n",
              " '/content/FOUND_RFQ_Head_Injury.csv',\n",
              " '/content/FOUND_RFQ_Height___Weight.csv',\n",
              " '/content/FOUND_RFQ_Occupation.csv',\n",
              " '/content/FOUND_RFQ_Pesticides_at_Work.csv',\n",
              " '/content/FOUND_RFQ_Pesticides_Non-Work.csv',\n",
              " '/content/FOUND_RFQ_Physical_Activity.csv',\n",
              " '/content/FOUND_RFQ_Residential_History.csv',\n",
              " '/content/FOUND_Self-Reported_Dx.csv',\n",
              " '/content/FOUND_RFQ_Smoking_History.csv',\n",
              " '/content/FOUND_RFQ_Toxicant_History.csv',\n",
              " '/content/FOUND_Enrollment_Status.csv',\n",
              " '/content/Family_History__PD_.csv',\n",
              " '/content/Features_of_REM_Behavior_Disorder.csv',\n",
              " '/content/Florbetaben_Eligibility.csv',\n",
              " '/content/Gait_Data___Arm_swing.csv',\n",
              " '/content/General_Medical_History.csv',\n",
              " '/content/General_Neurological_Exam.csv',\n",
              " '/content/General_Physical_Exam.csv',\n",
              " '/content/Genetic_and_Registry_Diagnostic_Questionnaire.csv',\n",
              " '/content/Genetic_Testing_Results.csv',\n",
              " '/content/Geriatric_Depression_Scale__Short_.csv',\n",
              " '/content/Hopkins_Verbal_Learning_Test.csv',\n",
              " '/content/IUSM_CATALOG.csv',\n",
              " '/content/IUSM_BIOSPECIMEN_CELL_CATALOG.csv',\n",
              " '/content/Inclusion_Exclusion.csv',\n",
              " '/content/Initiation_of_PD_Medication-_incidents.csv',\n",
              " '/content/Laboratory_Procedures.csv',\n",
              " '/content/Laboratory_Procedures_with_Elapsed_Times.csv',\n",
              " '/content/Letter_-_Number_Sequencing__PD_.csv',\n",
              " '/content/Lumbar_Puncture_Sample_Collection.csv',\n",
              " '/content/MDS_UPDRS_Part_I.csv',\n",
              " '/content/MDS_UPDRS_Part_I__Patient_Questionnaire.csv',\n",
              " '/content/MDS_UPDRS_Part_II__Patient_Questionnaire.csv',\n",
              " '/content/MDS_UPDRS_Part_III.csv',\n",
              " '/content/MDS_UPDRS_Part_IV.csv',\n",
              " '/content/MRI_Imaging_Data_Transfer_Information_Source_Document.csv',\n",
              " '/content/Magnetic_Resonance_Imaging.csv',\n",
              " '/content/Modified_Schwab_+_England_ADL.csv',\n",
              " '/content/Montreal_Cognitive_Assessment__MoCA_.csv',\n",
              " '/content/Neurological_Exam_-_Cranial_Nerves.csv',\n",
              " '/content/Olfactory_UPSIT.csv',\n",
              " '/content/PD_Features.csv',\n",
              " '/content/Page_Descriptions.csv',\n",
              " '/content/Patient_Status.csv',\n",
              " '/content/PASE_-_Household_Activity.csv',\n",
              " '/content/PASE_-_Leisure_Time_Activity.csv',\n",
              " '/content/Pilot_Biospecimen_Analysis_Results.csv',\n",
              " '/content/Pregnancy_Form.csv',\n",
              " '/content/Primary_Diagnosis.csv',\n",
              " '/content/Prodromal_Diagnostic_Questionnaire.csv',\n",
              " '/content/QUIP_Current_Short.csv',\n",
              " '/content/RBD_PSG_Eligibility.csv',\n",
              " '/content/REM_Sleep_Disorder_Questionnaire.csv',\n",
              " '/content/Research_Advance_Directive.csv',\n",
              " '/content/SCOPA-AUT.csv',\n",
              " '/content/SPECT_Scan_Information_Source_Document.csv',\n",
              " '/content/Screening___Demographics.csv',\n",
              " '/content/Semantic_Fluency.csv',\n",
              " '/content/Signature_Form.csv',\n",
              " '/content/Skin_Biopsy.csv',\n",
              " '/content/Skin_Biopsy_Eligibility.csv',\n",
              " '/content/Socio-Economics.csv',\n",
              " '/content/State-Trait_Anxiety_Inventory.csv',\n",
              " '/content/Surgery_for_Parkinson_Disease.csv',\n",
              " '/content/Symbol_Digit_Modalities.csv',\n",
              " '/content/ST_CATALOG.csv',\n",
              " '/content/TAP-PD_OPDM_Assessment.csv',\n",
              " '/content/TAP-PD_OPDM_Use_Questionnaire.csv',\n",
              " '/content/TAP-PD_Conclusion.csv',\n",
              " '/content/TAP-PD_Kinetics_Device_Testing.csv',\n",
              " '/content/TAP-PD_Subject_Eligibility.csv',\n",
              " '/content/University_of_Pennsylvania_Smell_ID_Test.csv',\n",
              " '/content/Use_of_PD_Medication.csv',\n",
              " '/content/Vital_Signs.csv',\n",
              " '/content/Whole_Blood_Sample_Collection.csv',\n",
              " '/content/iPSC_Blood_Sample.csv']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sHz0bYKQyHLK"
      },
      "source": [
        "A **Dictionary** is created with important covariates which are to be extracted from downloaded **.csv** files. Each **.csv** file represents different pre-clinical assessment test conducted to access patients condition against biomarkers of Parkinson's Disease ."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aVrnkiVZDTPB"
      },
      "source": [
        "'''\n",
        "\n",
        "PATNO: Unique ID assigned to each patient\n",
        "\n",
        "EVENT_ID: Patients Visit ID for recurring assessment carried out every month\n",
        "{BL(Baseline), V01(visit 1), V02, V03, V04, V05, V06, V07, V08, V09, V10, V11, V12}\n",
        "\n",
        "(PATNO + EVENT_ID) is unique to each record in all motor and non-motor assessment tests.\n",
        "\n",
        "'''\n",
        "covariates = {}\n",
        "\n",
        "########## SUBJECT CHARACHTERISTIC AND MEDICAL HISTORY ##########\n",
        "\n",
        "#Patients enrolment status\n",
        "covariates[\"pat_status\"] = [\"PATNO\", \"RECRUITMENT_CAT\", \"IMAGING_CAT\", \"ENROLL_DATE\", \"ENROLL_CAT\"]\n",
        "#Demographic infomation of patients \n",
        "covariates[\"pat_demographics\"] = [\"PATNO\", \"BIRTHDT\", \"GENDER\", \"APPRDX\", \"CURRENT_APPRDX\", \"HISPLAT\", \"RAINDALS\", \"RAASIAN\", \"RABLACK\", \"RAHAWOPI\", \"RAWHITE\", \"RANOS\"]\n",
        "#Family History Table Covariates\n",
        "covariates[\"pat_family_history\"] = [\"PATNO\", \"BIOMOMPD\", \"BIODADPD\", \"FULSIBPD\", \"HAFSIBPD\", \"MAGPARPD\", \"PAGPARPD\", \"MATAUPD\", \"PATAUPD\", \"KIDSPD\"]\n",
        "\n",
        "########## MOTOR AND NON-MOTOR ASSESSMENTS ##########\n",
        "\n",
        "#UPDRS - Unified Parkinson’s Disease Rating Scale (Part 1 - Part 4)\n",
        "covariates[\"updrs1\"] = [\"PATNO\", \"EVENT_ID\", \"INFODT\", \"NP1COG\", \"NP1HALL\", \"NP1DPRS\", \"NP1ANXS\", \"NP1APAT\", \"NP1DDS\"]\n",
        "covariates[\"updrs1pq\"] = [\"PATNO\", \"EVENT_ID\", \"NP1SLPN\", \"NP1SLPD\", \"NP1PAIN\", \"NP1URIN\", \"NP1CNST\", \"NP1LTHD\", \"NP1FATG\"]\n",
        "covariates[\"updrs2pq\"] = [\"PATNO\", \"EVENT_ID\", \"NP2SPCH\", \"NP2SALV\", \"NP2SWAL\", \"NP2EAT\", \"NP2DRES\", \"NP2HYGN\", \"NP2HWRT\", \"NP2HOBB\", \"NP2TURN\", \"NP2TRMR\", \"NP2RISE\", \"NP2WALK\", \"NP2FREZ\"]\n",
        "covariates[\"updrs3\"] = [\"PATNO\", \"EVENT_ID\", \"PAG_NAME\", \"EXAMTM\", \"NP3SPCH\", \"NP3FACXP\", \"NP3RIGN\", \"NP3RIGRU\", \"NP3RIGLU\", \"PN3RIGRL\", \"NP3RIGLL\", \"NP3FTAPR\", \"NP3FTAPL\", \"NP3HMOVR\", \"NP3HMOVL\", \"NP3PRSPR\", \"NP3PRSPL\", \"NP3TTAPR\", \"NP3TTAPL\", \"NP3LGAGR\", \"NP3LGAGL\", \"NP3RISNG\", \"NP3GAIT\", \"NP3FRZGT\", \"NP3PSTBL\", \"NP3POSTR\", \"NP3BRADY\", \"NP3PTRMR\", \"NP3PTRML\", \"NP3KTRMR\", \"NP3KTRML\", \"NP3RTARU\", \"NP3RTALU\", \"NP3RTARL\", \"NP3RTALL\", \"NP3RTALJ\", \"NP3RTCON\", \"DYSKPRES\", \"DYSKIRAT\", \"NHY\", \"ANNUAL_TIME_BTW_DOSE_NUPDRS\", \"ON_OFF_DOSE\", \"PD_MED_USE\"]\n",
        "# MOCA - Montreal Cognitive Assessment\n",
        "covariates[\"moca\"] = [\"PATNO\", \"EVENT_ID\", \"MCAALTTM\", \"MCACUBE\", \"MCACLCKC\", \"MCACLCKN\", \"MCACLCKH\", \"MCALION\", \"MCARHINO\", \"MCACAMEL\", \"MCAFDS\", \"MCABDS\", \"MCAVIGIL\", \"MCASER7\", \"MCASNTNC\", \"MCAVFNUM\", \"MCAVF\", \"MCAABSTR\", \"MCAREC1\", \"MCAREC2\", \"MCAREC3\", \"MCAREC4\", \"MCAREC5\", \"MCADATE\", \"MCAMONTH\", \"MCAYR\", \"MCADAY\", \"MCAPLACE\", \"MCACITY\", \"MCATOT\"]\n",
        "# STAI - State-Trait Anxiety Inventory Test\n",
        "covariates[\"stai\"] = [ \"PATNO\", \"EVENT_ID\", \"STAIAD1\", \"STAIAD2\", \"STAIAD3\", \"STAIAD4\", \"STAIAD5\", \"STAIAD6\", \"STAIAD7\", \"STAIAD8\", \"STAIAD9\", \"STAIAD10\", \"STAIAD11\", \"STAIAD12\", \"STAIAD13\", \"STAIAD14\", \"STAIAD15\", \"STAIAD16\", \"STAIAD17\", \"STAIAD18\", \"STAIAD19\", \"STAIAD20\", \"STAIAD21\", \"STAIAD22\", \"STAIAD23\", \"STAIAD24\", \"STAIAD25\", \"STAIAD26\", \"STAIAD27\", \"STAIAD28\", \"STAIAD29\", \"STAIAD30\", \"STAIAD31\", \"STAIAD32\", \"STAIAD33\", \"STAIAD34\", \"STAIAD35\", \"STAIAD36\", \"STAIAD37\", \"STAIAD38\", \"STAIAD39\", \"STAIAD40\" ]\n",
        "# GDS - Geriatric Depression Scale\n",
        "covariates[\"gds\"] = [\"PATNO\", \"EVENT_ID\", \"GDSSATIS\", \"GDSDROPD\", \"GDSEMPTY\", \"GDSBORED\", \"GDSGSPIR\", \"GDSAFRAD\", \"GDSHAPPY\", \"GDSHLPLS\", \"GDSHOME\", \"GDSMEMRY\", \"GDSALIVE\", \"GDSWRTLS\", \"GDSENRGY\", \"GDSHOPLS\", \"GDSBETER\"]\n",
        "# QUIP- Questionnaire for Impulsive-compulsive Disorder in Parkinson’s Disease \n",
        "covariates[\"quip\"] = [ \"PATNO\", \"EVENT_ID\", \"TMGAMBLE\", \"CNTRLGMB\", \"TMSEX\", \"CNTRLSEX\", \"TMBUY\", \"CNTRLBUY\", \"TMEAT\", \"CNTRLEAT\", \"TMTORACT\", \"TMTMTACT\", \"TMTRWD\" ]\n",
        "# SCOPA-AUT - Scale for Outcomes in Parkinson’s Disease – Autonomic Dysfunction Test \n",
        "covariates[\"scopa_aut\"] = [ \"PATNO\", \"EVENT_ID\", \"SCAU1\", \"SCAU2\", \"SCAU3\", \"SCAU4\", \"SCAU5\", \"SCAU6\", \"SCAU7\", \"SCAU8\", \"SCAU9\", \"SCAU10\", \"SCAU11\", \"SCAU12\", \"SCAU13\", \"SCAU14\", \"SCAU15\", \"SCAU16\", \"SCAU17\", \"SCAU18\", \"SCAU19\", \"SCAU20\", \"SCAU21\", \"SCAU22\", \"SCAU23\", \"SCAU23A\", \"SCAU23AT\", \"SCAU24\", \"SCAU25\", \"SCAU26A\", \"SCAU26AT\", \"SCAU26B\", \"SCAU26BT\", \"SCAU26C\", \"SCAU26CT\", \"SCAU26D\", \"SCAU26DT\" ]\n",
        "# SFT - Semantic Fluency Test\n",
        "covariates[\"sft\"] = [ \"PATNO\", \"EVENT_ID\", \"VLTANIM\", \"VLTVEG\", \"VLTFRUIT\" ]\n",
        "# REM-RBD - Rapid Eye Movement Sleep Behavior Disorder Test\n",
        "covariates[\"rem_rbd\"] = [ \"PATNO\", \"EVENT_ID\", \"DRMVIVID\", \"DRMAGRAC\", \"DRMNOCTB\", \"SLPLMBMV\", \"SLPINJUR\", \"DRMVERBL\", \"DRMFIGHT\", \"DRMUMV\", \"DRMOBJFL\", \"MVAWAKEN\", \"DRMREMEM\", \"SLPDSTRB\", \"STROKE\", \"HETRA\", \"RLS\", \"NARCLPSY\", \"DEPRS\", \"EPILEPSY\", \"BRNINFM\" ]\n",
        "#EPWORTH - Epworth Sleepness Scale\n",
        "covariates[\"epworth\"] = [\"PATNO\", \"EVENT_ID\", \"ESS1\", \"ESS2\", \"ESS3\", \"ESS4\", \"ESS5\", \"ESS6\", \"ESS7\", \"ESS8\"]\n",
        "# HVLT- Hopkins Verbal Learning Test\n",
        "covariates[\"hvlt\"] = [\"PATNO\", \"EVENT_ID\", \"HVLTRT1\", \"HVLTRT2\", \"HVLTRT3\", \"HVLTRDLY\", \"HVLTREC\", \"HVLTFPRL\", \"HVLTFPUN\"]\n",
        "# LNS - Letter Number Sequencing Test\n",
        "covariates[\"lns\"] = [\"PATNO\", \"EVENT_ID\",\"LNS_TOTRAW\"]\n",
        "# Neurological exam: Cranial Nerves Test\n",
        "covariates[\"neuro_cranial\"] = [\"PATNO\", \"EVENT_ID\",\"CN1RSP\", \"CN2RSP\", \"CN346RSP\", \"CN5RSP\", \"CN7RSP\", \"CN8RSP\", \"CN910RSP\", \"CN11RSP\", \"CN12RSP\"]\n",
        "# SDM - Symbol Digit Modalities Test\n",
        "covariates[\"sdm\"] = [ \"PATNO\", \"EVENT_ID\", \"SDMTOTAL\"] \n",
        "# Neuropsychological Test - (Line Orientation)\n",
        "covariates[\"benton\"] = [\"PATNO\", \"EVENT_ID\", \"JLO_TOTRAW\"]\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eCz4h9Fn2S5q"
      },
      "source": [
        "Extracting required data from **.csv** files into dataframes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4OMyUCdKkMkD"
      },
      "source": [
        "path=''\n",
        "\n",
        "########## SUBJECT CHARACHTERISTIC AND MEDICAL HISTORY ##########\n",
        "pat_status = pd.read_csv(path+\"Patient_Status.csv\", index_col=[\"PATNO\"], usecols=covariates[\"pat_status\"])\n",
        "pat_demographics = pd.read_csv(path+\"Screening___Demographics.csv\", index_col=[\"PATNO\"], usecols=covariates[\"pat_demographics\"])\n",
        "pat_family_history = pd.read_csv(path+\"Family_History__PD_.csv\", index_col=[\"PATNO\"], usecols=covariates[\"pat_family_history\"])\n",
        "\n",
        "########## MOTOR AND NON-MOTOR ASSESSMENTS ##########\n",
        "updrs1 = pd.read_csv(path+\"MDS_UPDRS_Part_I.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], parse_dates=[\"INFODT\"], usecols=covariates[\"updrs1\"])\n",
        "updrs1pq = pd.read_csv(path+\"MDS_UPDRS_Part_I__Patient_Questionnaire.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"updrs1pq\"])\n",
        "updrs2pq = pd.read_csv(path+\"MDS_UPDRS_Part_II__Patient_Questionnaire.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"updrs2pq\"])\n",
        "updrs3_temp = pd.read_csv(path+\"MDS_UPDRS_Part_III.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"updrs3\"])\n",
        "updrs3 = updrs3_temp[updrs3_temp.PAG_NAME == 'NUPDRS3'] # pre med dose \n",
        "updrs3a = updrs3_temp[updrs3_temp.PAG_NAME == 'NUPDRS3A'] # post med dose  \n",
        "moca = pd.read_csv(path+\"Montreal_Cognitive_Assessment__MoCA_.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"moca\"])\n",
        "stai = pd.read_csv(path+\"State-Trait_Anxiety_Inventory.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"stai\"])\n",
        "gds = pd.read_csv(path+\"Geriatric_Depression_Scale__Short_.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"gds\"])\n",
        "quip = pd.read_csv(path+\"QUIP_Current_Short.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"quip\"])\n",
        "scopa_aut = pd.read_csv(path+\"SCOPA-AUT.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"scopa_aut\"])\n",
        "sft = pd.read_csv(path+\"Semantic_Fluency.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"sft\"])\n",
        "rem_rbd = pd.read_csv(path+\"REM_Sleep_Disorder_Questionnaire.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"rem_rbd\"])\n",
        "epworth = pd.read_csv(path+\"Epworth_Sleepiness_Scale.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"epworth\"])\n",
        "hvlt = pd.read_csv(path+\"Hopkins_Verbal_Learning_Test.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"hvlt\"])\n",
        "lns = pd.read_csv(path+\"Letter_-_Number_Sequencing__PD_.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"lns\"])\n",
        "neuro_cranial = pd.read_csv(path+\"Neurological_Exam_-_Cranial_Nerves.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"neuro_cranial\"])\n",
        "sdm = pd.read_csv(path+\"Symbol_Digit_Modalities.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"sdm\"])\n",
        "benton = pd.read_csv(path+\"Benton_Judgment_of_Line_Orientation.csv\", index_col=[\"PATNO\", \"EVENT_ID\"], usecols=covariates[\"benton\"])\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ulkZDNJCtZ53"
      },
      "source": [
        "###Exploratory Data Analysis\n",
        "\n",
        "Detailed analysis of fetched data is conducted to obtain better understanding on characteristic and quality of fetched data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c_Wj7TaRfguV",
        "outputId": "868ed0f2-455d-452c-cbfc-1cbe1f22da79",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 638
        }
      },
      "source": [
        "list_of_datasets = ['updrs1', 'updrs1pq', 'updrs2pq', 'updrs3', 'moca', 'stai', 'gds', 'quip', 'scopa_aut', \n",
        "            'sft', 'rem_rbd', 'epworth', 'hvlt', 'lns', 'neuro_cranial', 'sdm', 'benton']\n",
        "\n",
        "visit_ids = ['BL', 'V01', 'V02', 'V03', 'V04', 'V05', 'V06', 'V07', 'V08', 'V09', 'V10', 'V11', 'V12']\n",
        "last_visit = visit_ids[-1] #V12\n",
        "\n",
        "\n",
        "init_dataset = eval(list_of_datasets[0]).reset_index()\n",
        "plot = init_dataset.reset_index().groupby('EVENT_ID').size().reset_index().rename(columns={0:list_of_datasets[0]})\n",
        "\n",
        "for dataset in list_of_datasets[1:]:    \n",
        "    dataset_idx = eval(dataset).reset_index()\n",
        "    temp_plot = dataset_idx.reset_index().groupby('EVENT_ID').size().reset_index().rename(columns={0:dataset})\n",
        "    \n",
        "    plot = plot.merge(temp_plot, on='EVENT_ID', how='outer')\n",
        "    \n",
        "plot = plot.set_index('EVENT_ID').loc[visit_ids].plot(kind='bar', title=\"Number of patients per visit for various type of assessments\",\n",
        "                                                      figsize=(20, 10))\n",
        "plot.set_ylabel(\"Number of patients\");\n",
        "plot.set_xlabel(\"Assessment Visit IDs\");"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Xg-GGhhS4D0v"
      },
      "source": [
        "From above visualisation, it can be inferred that as we move towards last visit of assessment, steep decline in number of participant can be seen. Also for various tests are not mandatory to be conducted during each visit. Therefore, based presented data instead of selecting each visit, specific visits with most test data is selected. That is BL, V02, V04, V06, V08, V10, V12"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Twadyd3HBfZB"
      },
      "source": [
        "Records of data for all the essential covariates identified in selected list of clinical assessment is merged over unique index key defined on patient id and visit id for the filtered list of patients."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5xew3hOohHFK",
        "outputId": "ba949282-dca8-4202-ad9e-114c33e60889",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "#Selecting visits of interest\n",
        "visit_ids_of_interest = ['BL', 'V02', 'V04', 'V06', 'V08', 'V10', 'V12']\n",
        "last_visit = visit_ids_of_interest[-1] #V12\n",
        "\n",
        "# selecting patient data who participated from BL to last_visit\n",
        "dataset_temp = eval(list_of_datasets[0]).reset_index()\n",
        "list_of_ptnno = dataset_temp[ dataset_temp.EVENT_ID == last_visit ]['PATNO']\n",
        "\n",
        "for dataset in list_of_datasets:\n",
        "    dataset_temp = eval(dataset).reset_index()\n",
        "    patno_temp = dataset_temp[ dataset_temp.EVENT_ID == last_visit ]['PATNO']\n",
        "    list_of_ptnno = list_of_ptnno[ list_of_ptnno.isin(patno_temp)]\n",
        "\n",
        "print(f'Total {list_of_ptnno.size} number of patients who participated till last visit for all assessments')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total 476 number of patients who participated till last visit for all assessments\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sWXdnjsBa-qo",
        "outputId": "13be54f7-8e21-414c-d24d-9b0093e7de96",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        }
      },
      "source": [
        "# constructing the data_visits dictionary with all the information\n",
        "merged_data = {}\n",
        "status = pat_status[pat_status.index.isin(list_of_ptnno)].ENROLL_CAT\n",
        "demographics = pat_demographics[pat_demographics.index.isin(list_of_ptnno)]\n",
        "merged_data[\"info\"] = pd.concat([status, demographics], axis=1)\n",
        "\n",
        "for dataset in list_of_datasets:    \n",
        "    dataset_temp = eval(dataset).reset_index()\n",
        "    merged_data[dataset] = dataset_temp[ dataset_temp['PATNO'].isin(list_of_ptnno) & dataset_temp['EVENT_ID'].isin(visit_ids_of_interest) ]\n",
        "\n",
        "merged_data[\"info\"].ENROLL_CAT.reset_index().groupby(\"ENROLL_CAT\").size()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ENROLL_CAT\n",
              "GENPD         8\n",
              "GENUN         6\n",
              "HC          154\n",
              "PD          294\n",
              "PRODROMA     10\n",
              "REGPD         1\n",
              "SWEDD         3\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fpxm1eHz6Csi",
        "outputId": "0e4a6f89-1ca6-47ba-d813-379f38cae481",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 354
        }
      },
      "source": [
        "plot_1 = merged_data[\"info\"].ENROLL_CAT.reset_index().groupby(\"ENROLL_CAT\").size().plot(kind='bar',\n",
        "                                                                                        title=\"Categories of selected participants\", \n",
        "                                                                                        color = ['r','r','g','g','r','r','r'])\n",
        "plot_1.set_ylabel(\"Number of Patients\"); \n",
        "plot_1.set_xlabel(\"Patient Category\");\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAFSCAYAAADhDegVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd1hT9+M98JOwxYFQFFCrrR9FXAUMuBXRqijugSLaahWctXVUWjfWbW3d1Fo3buvAUequtXVgtZU66gCrLJWhoOy8f3/4M18jEC5qkoue1/P0Kbn35t6TGHK4WyGEECAiIiqC0tgBiIioZGBhEBGRJCwMIiKShIVBRESSsDCIiEgSFgYREUnCwiBZcnNzw507dwy2vAcPHqBfv35wc3PDnDlzXtt8vb298fvvv7+2+b2MM2fOoEWLFkZbflxcHNzc3JCXl6dzusjISLRr185AqehlsDDeYOHh4ejevTvc3NzQrFkzDB48GJGRkZKe6+zsjNu3b+s5YeEuXLiAKlWqGGx5W7duRfny5fHnn38iODjYYMuVIjg4GN9++62xY0j2Ykk6OTnhwoULMDEx0fk8lUqFiIgIfcfD3bt34ezsjNzcXL0v601jauwApB9r1qzBypUrMX36dDRr1gxmZmY4efIkjhw5ApVKZex4hcrNzYWpqeE/lnFxcahevToUCoXBl/2mMNa/HRmQoDfOo0ePhKurqzhw4ECh0/z111+id+/eokGDBqJp06Zi+vTpIisrSwghhL+/v6hZs6b44IMPhKurq9i/f78QQoijR4+Kzp07iwYNGgg/Pz9x5coVzfyioqJEly5dhKurqxg1apQYPXq0WLhwoWb81q1bRZs2bYSHh4cICgoSCQkJmnE1a9YUGzduFB9++KFo1aqVZlhMTIwQQoisrCwxZ84c0bJlS9G4cWMxefJkkZGRIYQQIikpSQQGBooGDRoIDw8P0bdvX5GXl1fgaz5//rzo3r27cHd3F927dxfnz58XQggxYcIEUbt2bVGnTh3h6uoqTp06le+5x48fFz4+PsLV1VU0a9ZMrFq1SjNO1/vSqlUrzfzy8vLE999/L1q3bi08PT3Fp59+KlJSUjTTnjt3Tvj5+YkGDRqIFi1aiJ07d4otW7ZoZQsKChJCCJGQkCBGjhwpGjZsKFq1aiXWrVunmU9GRoaYMGGCUKlUwsfHR/zwww+iefPmhXwSnr7X69atE97e3sLT01PMmTNH8x7evn1b9O/fX3h6egpPT08xZswY8fDhQ63X9/333wtfX19Rp04d8fnnnwtnZ2dRr1494erqKlauXCnu3LkjatasKXJycoQQQqSkpIjg4GDRtGlToVKpxLBhw4QQQpw+fVorZ6tWrURoaKjw8fERKpVKBAcHi8zMTCGEEKmpqSIwMFA0bNhQqFQqERgYKOLj4zXPDQgIEN9++63w8/MTrq6uYuDAgSIpKUkIIUTLli1FzZo1haurq3B1dRV//vmniImJEf369RPu7u7C09NTjB49utD3623GwngDnThxQri4uGh+QQty6dIlceHCBZGTkyPu3Lkj2rdvL9asWaMZ//wXthBC/PPPP6JRo0bi4sWLIjc3V/z000+iVatWIisrS2RlZQkvLy+xdu1akZ2dLSIiIkSdOnU0hfH7778LT09PERUVJbKyskRISIjw9/fXWtbHH38sUlJSNEXw/PJnzpwpgoKCREpKikhLSxNBQUFiwYIFQgghFixYICZPniyys7NFdna2OHfunFCr1fleb0pKilCpVGLXrl0iJydHhIeHC5VKJZKTk4UQT0vj+YJ7UdOmTcW5c+eEEE+/rKKioop8X4TQLoy1a9eKXr16ifj4eJGVlSUmT54sPv/8cyGEEHfv3hWurq4iPDxcZGdni+TkZHH58uUCs+Xl5Ylu3bqJJUuWiKysLPHff/8Jb29v8euvvwohhJg/f77o27evSElJEXFxcaJjx45FFkZAQIBISUkRsbGxom3btmLbtm1CCCFiYmLEb7/9JrKyskRSUpLw9/cXX3/9tea5rVq1Ep07dxZxcXGaf7vnX7MQIl9hDBkyRIwePVqkpqaK7OxscebMGSFEwYXRsWNHERcXJ1JSUoSfn5/mfUhOThY///yzePLkiUhLSxOjRo3SFI8QTwujdevW4tatWyIjI0MEBASI+fPnF5hHCCE+//xzsXz5cpGXlycyMzM1/9akjfsw3kCpqakoX768zs0DdevWhaurK0xNTVG5cmX4+fnh3LlzhU6/detW+Pn54YMPPoCJiQm6desGMzMzXLx4EX/99Rdyc3MxYMAAmJmZoW3btqhXr57mueHh4ejRowfq1KkDc3NzjBkzBhcvXsTdu3c10wQGBsLGxgaWlpZayxVCYNu2bfjqq69gY2OD0qVLIygoCPv37wcAmJqa4v79+4iLi4OZmRlUKlWBm5WOHz+OqlWromvXrjA1NYWvry/ef/99HDt2TNJ7ampqihs3biA9PR3lypVDnTp1inxfXrRlyxZ8/vnncHBwgLm5OUaOHImIiAjk5uZi3759aNKkCXx9fWFmZoby5cvDxcWlwCyXLl1CcnIyRo4cCXNzc1SpUgW9e/fGgQMHAAAHDx7E0KFDYWNjA0dHR/Tv37/I1zdkyBDY2NjAyckJAwYMwL59+wAAVatWRdOmTWFubg5bW1sMHDgw3+ekf//+cHR0zPdvV5B79+7h119/xfTp01GuXDmYmZnB09Oz0On79esHR0dH2NjYYNiwYZp/9/Lly6Ndu3awsrJC6dKlMWzYsHy5unfvjvfeew+WlpZo3749rly5UuhyTE1NERcXh3v37sHCwkLWm22NiRsc30A2NjZISUnRuU05Ojoac+bMQVRUFDIyMpCXl6f5EixIXFwcdu/ejY0bN2qG5eTk4N69e1AoFKhYsaLWF7Wjo6Pm53v37mnN29raGjY2NkhMTETlypXzTf+85ORkZGRkoHv37pphQgio1WoAwCeffIKlS5di0KBBAAA/Pz8EBgbmm8+9e/fg5OSkNczJyQmJiYmFvubnLV68GCtWrMA333wDZ2dnjB07Fm5ubjrflxfFxcVhxIgRUCr/7+80pVKJpKQkxMfH491335WUJTY2Fvfu3dP6UsvLy9M8vnfvntb7+eLrLsjz01eqVEmT/8GDB5g5cyYiIyPx+PFjCCFQtmzZQp9blISEBJQrVw7lypWTNP2Lr+NZroyMDMyePRsnT57Ew4cPAQCPHz9GXl6eZue6vb295rlWVlZ48uRJocsZP348Fi1ahJ49e6JcuXIYOHAgevbsKfl1vS1YGG8gNzc3mJub4/Dhw2jfvn2B00ybNg21a9fGN998g9KlS2Pt2rU6j1BxdHTE0KFDMWzYsHzjzp49i8TERAghNKURHx+vOcqpQoUKiI2N1Uz/5MkTpKamomLFipphhe1sLl++PCwtLbF//36t6Z8pXbo0goODERwcjH///RcfffQR6tWrh8aNG2tNV6FCBcTFxWkNi4+PR/PmzQt9zc+rX78+VqxYgZycHISFheGzzz7DiRMndL4vL3JwcMCsWbPQoEGDfOMcHR3x999/F/i8F98bR0dHVK5cGb/88kuB09vb2yM+Ph41atQA8PR1FuX56ePi4lChQgUAwMKFC6FQKBAeHg4bGxscPnwYISEhOvPp4uDggIcPH+LRo0f5iqewXM88n2v16tWIjo7Gtm3bYG9vjytXrqBr164QEi6+XVBee3t7fP311wCeHt47cOBAeHh4oGrVqlJf2luBm6TeQGXKlMGnn36KkJAQHD58GBkZGcjJycGJEycwb948AE//GrO2toa1tTVu3ryJzZs3a83jnXfe0ToPolevXtiyZQv++usvCCHw5MkTHD9+HOnp6XB1dYWJiQk2btyI3NxcHD58GJcuXdI819fXFz/99BOuXLmC7OxsLFy4EPXr19esXeiiVCrRq1cvzJo1C0lJSQCAxMREnDx5EgBw7Ngx3L59G0IIlClTBiYmJgV+IbRs2RIxMTEIDw9Hbm4uDhw4gBs3bsDLy6vIDNnZ2di7dy/S0tJgZmYGa2trzVqCrvflRX379sV3332nKc/k5GQcPnwYANCpUyf8/vvvOHDgAHJzc5GSkqLZhGJnZ6e1+a5+/fqwtrbGypUrkZmZiby8PPz777+awvHx8cHKlSvx8OFDJCQkYMOGDUW+xh9//BEPHz5EfHw81q9fjw4dOgB4+jkpVaoUypQpg8TERKxatarIeb342XlehQoV0KJFC0yfPh0PHz5ETk6Ozk2hmzZtQkJCAlJTUxEaGqqVy8LCAmXLlkVqaiqWLl1aZK5nbG1toVQqtTIePHgQCQkJAIBy5cpBoVBorQnSU3xH3lCDBg1CcHAwli9fjsaNG8PLywthYWFo06YNAGDChAnYt28f3N3dMXnyZM0v4jMjR45EcHAwVCoVDhw4gHr16mHGjBkICQmBh4cH2rZti59++gkAYG5ujiVLlmDHjh3w8PDA3r174eXlBXNzcwBAkyZNMHr0aIwaNQrNmjXDnTt3inVewfjx41G1alX07t0b7u7u+PjjjxEdHQ0AuH37NgYOHAg3Nzf4+fmhb9++aNSoUb55lC9fHqGhoVizZg0aNmyIVatWITQ0FLa2tpIy7NmzB97e3nB3d8eWLVswf/58AND5vrxowIAB8Pb2xqBBg+Dm5obevXtrvuSdnJzwww8/YM2aNfD09ETXrl1x9epVAEDPnj1x48YNqFQqDB8+HCYmJggNDcXVq1fRunVrNGrUCJMmTdKU1MiRI+Hk5ITWrVtj0KBB6NKlS5Gvr3Xr1ujevTu6du0KLy8vzeaYkSNH4vLly1CpVAgMDETbtm2LnFdgYCBWrFgBlUqFH3/8Md/4efPmwdTUFD4+PmjSpAnWrVtX6Lx8fX0xaNAgtGnTBu+++65mTe6jjz5CVlYWGjVqBD8/P8lrisDTzVNDhw5F3759oVKpcPHiRVy6dAm9evWCm5sbhg0bhokTJxr0PKCSQiGkrMMRFVOvXr3Qp08f9OjRw9hRqAjOzs745ZdfZLf5xdvbG19//TWaNGli7Cj0/3ENg16Ls2fP4v79+8jNzcWuXbtw7dq1Yv3VR0Tyx53e9FpER0fjs88+Q0ZGBipXrozFixdrdlAS0ZuBm6SIiEgSbpIiIiJJWBhERCTJG78PIyXlMdRq/Wx1s7MrjaSk/MfblwQlOTvA/MbG/Malz/xKpQLly1sXOO6NLwy1WuitMJ7Nv6QqydkB5jc25jcuY+Q3WGEMHz4cd+/ehVKpRKlSpTB58mS4uLggOjoawcHBSE1NhY2NDebOnYtq1aoBgM5xRERkWAbbhzF37lzs3bsXu3fvxqBBg/DVV18BAKZOnQp/f39ERETA398fU6ZM0TxH1zgiIjIsgxVGmTJlND+np6dDoVAgKSkJly9fhq+vL4CnlwG4fPkykpOTdY4jIiLDM+g+jIkTJ+LUqVMQQmDVqlWIj49HxYoVNZcjNjExQYUKFRAfHw8hRKHjpF7/B3i6c0if7O3LFD2RTJXk7ADzGxvzG5cx8hu0MGbOnAkA2L17N+bNm4fRo0frfZlJSel62zlkb18G9++n6WXe+laSswPMb2zMb1z6zK9UKgr9Q9so52F07doVZ86cgYODAxITE5GXlwfg6U1gnt38xdHRsdBxRERkeAYpjMePH2vdCOXo0aMoV64c7Ozs4OLiorkd5L59++Di4gJbW1ud44iIyPAMskkqIyMDo0ePRkZGBpRKJcqVK4fQ0FAoFApMmzZNc9+GsmXLYu7cuZrn6RpHRESG9cZffJD7MApWkrMD8stfxsYClmbmept/Zk420lKz9Db/4pLb+19czF84Xfsw3vgzvYkMwdLMHL23Fn1f75e1zW8F0iCfwqC3Ey8+SEREkrAwiIhIEhYGERFJwsIgIiJJWBhERCQJC4OIiCRhYRARkSQsDCIikoSFQUREkrAwiIhIEhYGERFJwsIgIiJJWBhERCQJC4OIiCRhYRARkSQsDCIikoSFQUREkrAwiIhIEhYGERFJwsIgIiJJWBhERCQJC4OIiCRhYRARkSQsDCIikoSFQUREkrAwiIhIElNDLCQlJQVffPEF/vvvP5ibm6Nq1aoICQmBra0tnJ2dUbNmTSiVT7tr3rx5cHZ2BgAcPXoU8+bNQ15eHurUqYPZs2fDysrKEJGJiOgFBlnDUCgUGDx4MCIiIhAeHo4qVapgwYIFmvFbtmzBnj17sGfPHk1ZPH78GJMnT0ZoaCgOHToEa2tr/Pjjj4aIS0REBTBIYdjY2KBhw4aax66uroiLi9P5nF9//RV169ZFtWrVAAB9+vTBwYMH9RmTiIh0MMgmqeep1Wps3rwZ3t7emmH9+/dHXl4eWrRogVGjRsHc3Bzx8fFwcnLSTOPk5IT4+PhiL8/OrvRryV0Ye/syep2/PpXk7EDJz19ccnu9cstTXMxffAYvjBkzZqBUqVIICAgAABw/fhyOjo5IT0/H+PHjsWzZMnz++eevbXlJSelQq8Vrm9/z7O3L4P79NL3MW99KcnZAfvkN8csrt9crpzzFxfyFUyoVhf6hbdCjpObOnYvbt2/ju+++0+zkdnR0BACULl0avXr1wp9//qkZ/vxmq7i4OM20RERkeAYrjIULFyIqKgrLli2Dubk5AODhw4fIzMwEAOTm5iIiIgIuLi4AgObNm+PSpUuIiYkB8HTHuI+Pj6HiEhHRCwyySer69ev4/vvvUa1aNfTp0wcAULlyZQwePBhTpkyBQqFAbm4u3NzcMHr0aABP1zhCQkIQFBQEtVoNFxcXTJw40RBxiYioAAYpjBo1auDatWsFjgsPDy/0eW3atEGbNm30FYuIiIqBZ3oTEZEkLAwiIpKEhUFERJKwMIiISBIWBhERSfJShXHnzh3cvXv3dWchIiIZk1QYY8aM0ZyBvXPnTnTs2BG+vr7Yvn27XsMREZF8SCqMP/74A3Xr1gUArF27FmvWrMH27dvxww8/6DUcERHJh6QT93JycmBubo7ExESkpqaiQYMGAIAHDx7oNRwREcmHpMJwcXHB999/j9jYWHh5eQEAEhMTUbq0fi8dTkRE8iFpk9TMmTPx77//IisrC5999hkA4MKFC+jUqZNewxERkXxIWsNISUnBN998ozWsffv2Wjc4IiKiN5ukNYyBAwcWOHzw4MGvNQwREcmXzjUMtVoNIYTWf8/8999/MDEx0XtAIiKSB52FUbt2bSgUCs3Pz1MqlRg6dKj+khERkazoLIwjR45ACIH+/ftj48aNmuEKhQK2trawtLTUe0AiIpIHnYVRqVIlAMCxY8cMEoaIiORL0lFSqampWL16Na5cuYInT55ojQsLC9NLMCIikhdJhTF27FhkZ2fDx8cHVlZW+s5EREQyJKkwLly4gNOnT8Pc3FzfeYiISKYknYfh7OyMhIQEfWchIiIZk7SG0ahRIwwePBjdu3fHO++8ozWuZ8+eeglGRETyIqkwIiMjUbFiRZw6dUpruEKhYGEQEb0lJBXGhg0b9J2DiIhkTvItWlNSUrB7926sWrUKwNPLm3O/BhHR20NSYZw9exbt27dHeHg4li1bBgC4ffs2pk2bps9sREQkI5IKY9asWfjuu+/w448/wtT06VasDz74AH///bekhaSkpGDIkCFo164dOnXqhJEjRyI5ORkAcPHiRXTu3Bnt2rXDoEGDkJSUpHmernFERGRYkgojNjYWjRs3BgDNxQjNzMyQl5cnaSEKhQKDBw9GREQEwsPDUaVKFSxYsABqtRrjx4/HlClTEBERAZVKhQULFgCAznFERGR4kgqjevXqOHnypNaw33//HTVr1pS0EBsbGzRs2FDz2NXVFXFxcYiKioKFhQVUKhUAoE+fPvj5558BQOc4IiIyPElHSQUHByMoKAheXl7IzMzElClTcPToUSxfvrzYC1Sr1di8eTO8vb0RHx+vddc+W1tbqNVqpKam6hxnY2MjeXl2dvq977i9fRm9zl+fSnJ2oOTnLy65vV655Sku5i8+SYXh6uqKvXv3Yu/evejRowccHR2xY8cOODg4FHuBM2bMQKlSpRAQEIBDhw4V+/nFlZSUDrVaFD3hS7C3L4P799P0Mm99K8nZAfnlN8Qvr9xer5zyFBfzF06pVBT6h7akwgCAihUrYsiQIa8UZO7cubh9+zZCQ0OhVCrh6OiIuLg4zfjk5GQolUrY2NjoHEdERIZXaGFMnjwZM2bMAACMHz9es7P7RfPmzZO0oIULFyIqKgorV67UXMSwbt26yMzMRGRkJFQqFbZs2YL27dsXOY6IiAyv0MKoXLmy5ueqVau+0kKuX7+O77//HtWqVUOfPn0081+2bBnmzZuHqVOnIisrC5UqVcL8+fMBPL0FbGHjiIjI8AotjKCgIM3Pfn5+sLe3zzfN/fv3JS2kRo0auHbtWoHj3N3dER4eXuxxRERkWJIOq23Xrl2Bwzt27PhawxARkXxJKgwh8h9llJ6eXuh+DSIievPoPEqqZcuWUCgUyMrKgpeXl9a41NRUrmEQEb1FdBbG/PnzIYRAYGCg1tFQCoUCdnZ2eP/99/UekIiI5EFnYXh6egIATp8+DSsrK4MEIiIieZJ04p6VlRWuXLmCyMhIpKSkaO3TGD16tN7CERGRfEja6b1161b07dsXp0+fxg8//IB///0Xa9aswX///afvfEREJBOSCmPVqlVYtWoVli1bBktLSyxbtgyLFi3S3BuDiIjefJIKIykpSXOZcaVSCbVajZYtW+LYsWN6DUdERPIhaRXBwcEBd+7cQZUqVVCtWjUcOXIE5cuXh5mZmb7zERGRTEgqjMGDB+PWrVuoUqUKhg8fjtGjRyMnJwcTJ07Udz4iIpKJIgsjLS0Nzs7OmgsQtmzZEmfPnkVOTg6sra31HpCIiORB5z6M48ePo3nz5ujRowdatmyJ06dPAwDMzc1ZFkREbxmdhbFo0SKMGzcOFy5cwKefforvvvvOULmIiEhmdBbGnTt3EBAQACsrK/Tr1w+3b982VC4iIpIZnYWhVqs1P5uamiIvL0/vgYiISJ507vTOzMxEv379NI8fP36s9RgAwsLC9JOMiIhkRWdhzJw5U+txz5499RqGiIjkS2dhdOvWzVA5iIhI5iRdGoSIiIiFQUREkrAwiIhIkkILo3fv3pqfly5dapAwREQkX4UWRkxMDLKysgAAq1evNlggIiKSp0KPkmrdujXatWuHSpUqISsrK9/5F8/wPAwiordDoYUxe/ZsREZGIjY2FpcuXeI5GEREbzmd52GoVCqoVCrk5OS88jkZc+fORUREBGJjYxEeHo6aNWsCALy9vWFubg4LCwsAwLhx49C8eXMAwMWLFzFlyhRkZWWhUqVKmD9/Puzs7F4pB8lTGRsLWJqZF+s59vZlJE+bmZONtNSs4sYioudIuoFSz549cebMGezevRv37t1DhQoV0KVLFzRq1Ejyglq3bo0BAwYUuGlr8eLFmgJ5Rq1WY/z48Zg9ezZUKhWWL1+OBQsWYPbs2ZKXSSWHpZk5em8dprf5b/NbgTSwMIhehaTDardv347PPvsM9vb2+PDDD1GhQgWMHTsW27Ztk7wglUoFR0dHydNHRUXBwsJCcy/xPn364Oeff5b8fCIier0krWGsWrUKa9asQa1atTTDfHx88Omnn2odfvuyxo0bByEEGjRogDFjxqBs2bKIj4+Hk5OTZhpbW1uo1WqkpqbCxsZG8rzt7Eq/cj5dirNZRG5KcvaXUdJfr9zyyy1PcTF/8UkqjNTUVFSvXl1r2Pvvv4+HDx++coCwsDA4OjoiOzsbM2fOREhICBYsWPDK830mKSkdarV4bfN7nr19Gdy/n6aXeeub3LIb4sOvz9db0vMXl9w+P8XF/IVTKhWF/qEtaZOUu7s75syZg4yMDADAkydPMG/ePLi5ub1yuGebqczNzeHv748///xTMzwuLk4zXXJyMpRKZbHWLoiI6PWRVBjTp0/H1atXoVKp0KRJE3h4eODq1auYPn36Ky38yZMnSEt72pJCCBw4cAAuLi4AgLp16yIzMxORkZEAgC1btqB9+/avtDwiInp5kjZJVahQAWFhYUhISNAcJeXg4FCsBX399df45Zdf8ODBAwwcOBA2NjYIDQ3FqFGjkJeXB7VajerVq2Pq1KkAAKVSiXnz5mHq1Klah9USEZFxSCqMZxwcHIpdFM9MmjQJkyZNyjd89+7dhT7H3d0d4eHhL7U8IiJ6vXi1WiIikoSFQUREkhRZGGq1Gn/88Qeys7MNkYeIiGSqyMJQKpUYPnw4zM2Ld50fIiJ6s0jaJOXh4YGLFy/qOwsREcmYpKOknJycMGTIELRu3RoODg5QKBSacaNHj9ZbOCIikg9JhZGVlYU2bdoAABITE/UaiIiI5ElSYfCS4kREJPnEvZs3b+Lnn39GUlISpkyZglu3biE7O1vrCrZERPTmkrTT++DBg+jXrx8SExM1Z2Y/fvwYc+bM0Ws4IiKSD0lrGIsXL8batWtRq1YtHDx4EABQq1YtXL16Va/hiIhIPiStYSQnJ8PZ2RkANEdIKRQKraOliIjozSapMOrUqYM9e/ZoDdu/fz/q16+vl1BERCQ/kjZJTZw4EZ988gl27NiBJ0+e4JNPPkF0dDRWr16t73xERCQTkgqjevXqOHjwII4dOwYvLy84OjrCy8sL1tbW+s5HREQyIfmwWisrKzRo0ACVK1dGxYoVWRZERG8ZSYURFxeHcePG4a+//kLZsmXx6NEjfPDBB5g/fz4qVaqk74xERCQDknZ6T5gwAXXq1MG5c+fwxx9/4OzZs6hbty6Cg4P1nY+IiGRC0hrGP//8g9WrV8PMzAwAYG1tjXHjxqFhw4Z6DUdERPIhaQ3D1dUVf//9t9awqKgouLm56SUUERHJT6FrGIsWLdL8XKVKFQQGBsLLywsODg5ISEjAiRMn4Ovra5CQRERkfIUWRkJCgtbjtm3bAnh61re5uTk+/PBDZGVl6TcdERHJRqGFwUuaExHR8ySfh5GRkYHbt2/jyZMnWsPd3d1feygiIpIfSYWxe/duhISEwMzMDJaWlprhCoUCxxLG1+8AACAASURBVI8f11c2IiKSEUmFMX/+fCxZsgRNmzbVdx4iIpIpSYfVmpmZwdPT86UXMnfuXHh7e8PZ2Rn//vuvZnh0dDT8/PzQrl07+Pn5ISYmRtI4IiIyPEmFMXr0aMyZMwfJyckvtZDWrVsjLCws32VEpk6dCn9/f0RERMDf3x9TpkyRNI6IiAxPUmFUq1YNR48eRdOmTeHi4gIXFxfUqlULLi4ukhaiUqng6OioNSwpKQmXL1/WnMvh6+uLy5cvIzk5Wec4IiIyDkn7ML744gt06dIFHTp00Nrp/Sri4+NRsWJFmJiYAABMTExQoUIFxMfHQwhR6DhbW9tiLcfOrvRryVsYe/syep2/PpXk7C+jpL9eueWXW57iYv7ik1QYqampGD16dIm8JWtSUjrUaqGXedvbl8H9+2l6mbe+yS27IT78+ny9JT1/ccnt81NczF84pVJR6B/akjZJde/ePd8tWl+Vo6MjEhMTkZeXBwDIy8vDvXv34OjoqHMcEREZh6Q1jL///hthYWFYsWIF3nnnHa1xYWFhL7VgOzs7uLi4YN++fejSpQv27dsHFxcXzSYnXeOIiMjwJBVG79690bt375deyNdff41ffvkFDx48wMCBA2FjY4P9+/dj2rRpCA4OxvLly1G2bFnMnTtX8xxd44iIyPAkFUa3bt1eaSGTJk3CpEmT8g2vXr06tm/fXuBzdI0jIiLDk1QYO3bsKHRcz549X1sYIiKSL0mF8eIO7wcPHuDOnTtwc3NjYRARvSUkFcaGDRvyDduxYwdu3rz52gMREZE8STqstiDdu3fHzp07X2cWIiKSMUlrGGq1WutxRkYG9u7dizJlSvaZkkREJJ2kwqhdu3a+s7wrVqyIGTNm6CUUERHJj6TCOHLkiNZjKysrnkRHRPSWkVQYL16WnIiI3j46C6N///46LzioUCiwbt261x6KiIjkR2dhdO7cucDhiYmJ2LBhAzIzM/USioiI5EdnYfTq1UvrcUpKClauXIlt27ahQ4cOGDFihF7DERGRfEjah5Geno5Vq1YhLCwMXl5e2LVrF9599119ZyMiIhnRWRiZmZlYt24dVq9ejYYNG2LTpk2oUaOGobIREZGM6CwMb29vqNVqDB48GHXr1sWDBw/w4MEDrWkaN26s14BERCQPOgvj2f27N2/eXOB4hUKR7xwNIiJ6M+ksjKNHjxoqBxERydxLX3yQiIjeLiwMIiKShIVBRESSsDCIiEgSFgYREUnCwiAiIklYGEREJAkLg4iIJGFhEBGRJCwMIiKSRNLlzfXN29sb5ubmsLCwAACMGzcOzZs3x8WLFzFlyhRkZWWhUqVKmD9/Puzs7Iyclojo7SSLwgCAxYsXo2bNmprHarUa48ePx+zZs6FSqbB8+XIsWLAAs2fPNmJKIqK3l2w3SUVFRcHCwgIqlQoA0KdPH/z8889GTkVE9PaSzRrGuHHjIIRAgwYNMGbMGMTHx8PJyUkz3tbWFmq1GqmpqbCxsZE8Xzu70vqIq2FvX0av89enkpz9ZZT01yu3/HLLU1zMX3yyKIywsDA4OjoiOzsbM2fOREhICD788MPXMu+kpHSo1eK1zOtF9vZlcP9+ml7mrW9yy26ID78+X29Jz19ccvv8FBfzF06pVBT6h7YsNkk5OjoCAMzNzeHv748///wTjo6OiIuL00yTnJwMpVJZrLULIiJ6fYxeGE+ePEFa2tOmFELgwIEDcHFxQd26dZGZmYnIyEgAwJYtW9C+fXtjRiUieqsZfZNUUlISRo0ahby8PKjValSvXh1Tp06FUqnEvHnzMHXqVK3DaomIyDiMXhhVqlTB7t27Cxzn7u6O8PBwAyciIqKCGH2TFBERlQwsDCIikoSFQUREkrAwiIhIEhYGERFJwsIgIiJJWBhERCQJC4OIiCRhYRARkSQsDCIikoSFQUREkrAwiIhIEhYGERFJwsIgIiJJWBhERCQJC4OIiCRhYRARkSQsDCIikoSFQUREkrAwiIhIEhYGERFJwsIgIiJJWBhERCSJqbEDEJHxlS9jDlNLi2I9x96+jORpczOzkJKWXdxYJDMsjOfwl4beVqaWFjjVpYfe5t90z06An/0Sj4XxnJL8S8OyIyJ9Y2G8IUpy2RFRySD7nd7R0dHw8/NDu3bt4Ofnh5iYGGNHIiJ6K8l+DWPq1Knw9/dHly5dsGfPHkyZMgXr1683diwiotempGxSlnVhJCUl4fLly1izZg0AwNfXFzNmzEBycjJsbW0lzUOpVBRrmRYV7IudsziKm6c4SnJ2ALAvJe3f9GUxv24l+fNTrrQZTC30+IWblYWH6TnFjSWZqaUFIocM1dv8VT+EQvlYWn5d/04KIYR4XaFet6ioKEyYMAH79+/XDOvQoQPmz5+POnXqGDEZEdHbR/b7MIiISB5kXRiOjo5ITExEXl4eACAvLw/37t2Do6OjkZMREb19ZF0YdnZ2cHFxwb59+wAA+/btg4uLi+T9F0RE9PrIeh8GANy8eRPBwcF49OgRypYti7lz5+L99983diwioreO7AuDiIjkQdabpIiISD5YGEREJAkLg4iIJGFhEBGRJCwMIiKShIVBRESSyPrig3KTnp6O6OhoAMD7778Pa2trIyeSJiMjQ+d4KysrAyUhopchl+8enochgVqtxqxZs7BlyxZYWlpCCIGsrCz4+/vjyy+/hEKh36uIvqpatWpBoVDg+X/qZ5kVCgUuX75srGiSXL9+Hffv30eTJk20hv/++++oUKEC/ve//xkpmXRCCPz666+4ceMGAKBGjRpo3ry57D87umRlZeHgwYPo2rWrsaMUKT09HXv37tV6/zt16oTSpUsbOZlucvvu4SYpCTZu3IioqCjs3bsXkZGROH/+PPbu3YuoqChs3LjR2PGKdPXqVVy5cgVXr17V/Dx79mxUqlQJbdq0MXa8In3zzTcFXg7Gzs4OCxYsMEKi4nn06BF69OiBadOm4fz584iMjMTUqVPRo0cPpKWlGTtesf3111+YMmUKmjVrht27dxs7TpESExPRqVMn7N27FyYmJlAqldizZw86deqExMREY8fTSXbfPYKK1LNnT3Hnzp18w+/cuSN69uxphEQv78iRI6JTp05i4MCB4tKlS8aOI0mPHj0KHefr62vAJC9n9uzZYtKkSSInJ0czLDs7W0yePFnMnj3biMmkS0pKEqtWrRIdO3YUbdu2FZ6eniIhIcHYsSSZPHmyWLRoUb7hS5YsEZMmTTJCIunk9t3DfRgSPHr0CJUrV843vHLlyiXmL8TIyEgsWLAAeXl5+PLLL9G4cWNjR5JM13uck6O/m9q8LqdOncLWrVthavp/v25mZmYIDg5G7969jZhMmhEjRuD8+fP48MMPERISAnd3d3h7e6NixYrGjiZJZGQk9u7dm294UFAQOnfubIRE0sntu4eFIYGuHUylSpUyYJKXExgYiFu3bmHkyJFo164dAO0d4XLf6W1ra4vLly+jdu3aWsMvX74MGxsbI6WSTghR4OekJHx2AODvv/9G5cqV4erqChcXFwAoUfteTExMtMr6GTMzswKHy4ncvnvk/W7JREJCAubNm5dvuBBC9ttAAeDXX38FAAQHB+PLL7/UDBdCQKFQ4MqVK8aKJsnw4cMxbNgwjBw5EvXq1QMAXLp0CcuXL8f06dONnK5o5ubmhY4zMzMzYJKXc+LECZw8eRI7d+7EvHnz0KpVK2RlZRk7lmS6SkHuhSG37x4eJSXB0qVLdY4fOXKkgZK8vX777TcsX74cly9fhkKhQO3atTF06FA0b97c2NGK5OrqWuCRXEII3Lp1CxcuXDBCqpeTkpKCPXv24KeffkJ6ejp8fX0xZswYY8fSqU6dOihbtmy+4UIIpKenIyoqygippJHbdw8Lg2Tv3Llzmp+ffVyf3yTi4eFh8EzFcfbsWaSkpCA2NhZVq1ZFmTJltMZ7enoaKdmr+fvvv/HTTz9h2rRpxo6iU2xsrM7xlSpVMlCSko+FIcHVq1fx1VdfITo6GrVr18acOXNQpUoVY8eSzNvbW+sLVqFQwM7ODk2bNsXQoUN1bjKRgx49emh+vnXrFqpXr65VHDt27DBWNEkOHDiAL7/8EtbW1sjOzsaSJUtK1EEHb8KJn9euXUNMTAycnZ1RrVo1Y8cpltjYWGzatEnrHBJ/f384OTkZPAsLQ4I+ffqgQ4cOaNq0KQ4ePIhr165hyZIlxo4l2fXr1/MNS0lJwdatW1G+fHlMmjTJCKleTteuXUvEsf/P69SpE+bNmwcXFxecPn0ay5Ytw4YNG4wdS7JnJ34CwItfFyVhH9j69euxePFivPfee4iOjkZISAg6dOhg7FiS3Lx5E3379kWzZs1Qv359AE/33/3222/YtGkTqlevbthABj+QtwTq1KmT1uPOnTsbKcnrlZubW+JeS9euXY0dodhefI+7dOlipCQvJyAgQPTo0UNs3LhRpKamGjtOsfn4+Ij4+HghhBDXr18Xfn5+Rk4k3ZgxY8TmzZvzDd+6dasYM2aMwfPI+xABmTAxMdF6rFS+GSfIPzvrlfQrJycHN2/e1Px1np2drfVY7pc22bBhA+7cuYPdu3fDz88PNWvWRI8ePdC8efMS8fkxNzeHg4MDgKfvdUk6wuuff/7BN998k294r169sHr1aoPnYWFI8O+//2ptc3706BEaN26sOSz1jz/+MGK6l3fq1Kl8O2Dl6Nm2W+Dp9Yue/7IF5P+Fm5mZiSFDhmgNe/ZYoVDgyJEjxohVLFWqVMGoUaMwatQoHD58GMHBwRg8eDA++eQTY0crUnp6Ok6cOFHo45YtWxojliSF7V9UKBRG2ffIwpDgl19+MXaEV9KjR498J1qlpKRAoVAUedieHAQGBmo9fv7LtyR84R49etTYEV6ZEAInT57Erl27cO3aNfj7+8PX19fYsSRxdHTEqlWrNI8dHBw0jxUKhawLQ6FQIDMzM9++o2fjDJ5HFJSE3ihnz57VeqxQKGBra4uqVavK/sQlMr758+fj8OHDcHd3R7du3UrsYcAlUWFXmhZGOumWhSHBp59+WmibKxQKfPfddwZORGQ4tWrVwnvvvYdSpUoV+Hsg98OahRDIzs6GhYUFAODixYuaa5DVrl27xNzXRg7456UErVq1yjcsLS0N69atQ0pKihESFQ8Lj17F+vXrjR3hlXz33XdQKBT47LPPAACfffYZ7OzskJ2djbZt22LUqFFGTli43NzcQrcC3L9/H/b29gbNwzWMYsrOzsb69euxdu1atGvXDsOHD4ednZ2xY+m0a9eufMOeL7w///zTCKmIDKNr167YtGmT5mJ9z87lycvLQ0BAADZv3mzkhIXr1q2b5vd3+PDhWL58eYHjDIVrGBKp1Wps374dK1asgKenJ7Zs2VLgZYflqFu3bpqfCyo8oqJERkZi6dKluHbtGgDA2dkZI0eOhEqlMnKyoimVSq0ruz47ac/ExET2l8d//u/5uLi4QscZCgtDgoMHD2LRokV477338MMPP6BGjRrGjlRsJbnwyLgOHz6MGTNmYOjQoZgwYQIA4MKFCxg7diwmT54s+7s2Pnr0SOvx80fdyX2T8ouX9ClsnKGwMCT4/PPP4eTkBDMzswIPQ120aJERUkn3JhQeGc/y5cuxatUqrc+Ni4sLVCoVJkyYIPvCqF27Nnbv3p3v3uN79+7Nd48VuVGr1ZrDap//+dk4Q+M+DAl++uknnW3+/CYfOapVqxacnJxQt27dAl+H3AuPjKtDhw44cOBAscfJRUxMDAICAtCwYUN88MEHAJ5eaff06dPYuHGjrC9GKLfDarmGIUH37t2NHeGVzJo1q0TdIY3kJScnBzk5Oflu9pSdnY3s7GwjpZKuWrVq2LVrF8LCwvDbb78BeLqG9NNPP6FChQpGTqfb1atXjR1BC9cwJFi4cKHmJjE7duxAz549NeMmT56MGTNmGCsakd7NmTMH9+7dw/Tp0zWXknn06BGmTZsGe3t7rbs40uu1c+dONGzYUDb7G+V/5TAZOHnypObnsLAwrXFyvlvXMwsXLtT8/OJJVpMnTzZ0HCphxowZA0tLS7Rs2RLdunVDt27d4OXlBUtLS9nfbQ94emviZ54/LBUAgoKCDB2nWA4dOoRu3brB29sbwcHB2LVrV5E3hNInbpKS4PmVsBdXyErCCtrJkyc1v9hhYWFaa0glofDIuMzNzTFr1iyMGDEC169fhxACNWvWLDF3qnt2KDDw9Av4+UPJjXFf7OIIDQ2FWq3GP//8g3PnziEiIgKzZ89GmTJl0LBhQ8yaNcugebiGIYHcDm0rrpJeeCQPlSpVgpeXF1q1aoVKlSppzumRO12f/5Lw+6tUKlGvXj0MGjQI48aNw6hRo2BqaoqDBw8aPAvXMCS4e/cuRo8ene9nIYRRVw+lKumFR8aVmZmJjRs3Ij4+Hu3atYOnpyc2b96MZcuW4X//+x8GDBhg7Ig66fr8y93Nmzdx5swZnDlzBlevXkW1atWgUqkwZ84c1KtXz+B5uNNbgqJOv5f7YbUeHh5o0qQJAOD333/X/CyEwOnTp/NdzZboeWPHjkVCQgLc3Nxw7tw5VKhQAdeuXcPEiRNlfWnwZ1xdXTX3TLlx44bmZyEEbt26hQsXLhgznk61atWCq6srhg0bhhYtWhi98FgYEty8eRPR0dGaE5RmzZqFtLQ0AMCAAQPg4uJizHhFKumFR8bl4+OD8PBwmJqaIj09Hc2aNcORI0dkfw21Z3T9QfTgwQNZ39/72LFjOHfuHM6dO4fMzEy4u7vD09MTnp6eBr/wIMBNUpIsXrxY61yMEydOYMCAAXjy5AlWrlyJb7/91ojpila/fn2dhUeki6WlpeaKqaVLl0a1atVKTFkAgKenJ+7fv4/ExETUqlULpqamSE5ORmhoKHbt2iXrwmjVqpXmatmPHz/G+fPnce7cOSxevBgKhQI///yzQfOwMCS4ffu21qq3lZUV+vXrBwCa/8tZSS88Mq7ExETMmzdP8/jevXtaj7/44gtjxJJs+/btmD59OsqVKwdbW1uMHj0awcHBaNasGXbu3GnseJIkJyfjzJkzOHv2LM6cOYOEhATUr1/f4DlYGBLk5eVpPX7+puwvXthMjkp64ZFx+fv763wsd2vXrsWuXbtQo0YNnD9/HgMGDMA333yD9u3bGztakaZNm4azZ88iNjYW9erVg6enJ6ZOnQo3Nzfe01uucnJykJ6ejtKlSwMAqlevDuDpzeRLwqURSnrhkXGNHDnS2BFeiampqebCiQ0aNECVKlVKRFkAgI2NDSZPngx3d3dYWFjg0aNHOHv2LG7duoVatWoZPA8LQ4KOHTviq6++wqxZszSlkZ6ejkmTJsl6++czJb3wyPhiY2OxadMm3LhxAwBQo0YN+Pv7w8nJycjJipaTk4ObN29qzsFQKpVaj58dNSVHd+/eRfv27WFhYYHU1FR06dIFpUuXRkpKCj7//HP06tXLoHl4lJQEubm5CA4OxpEjRzRXtoyJiUHr1q0xZ86cQm+hKBdLlizB9evXCyy89957T3NeCVFBbt68ib59+6JZs2aa7eaXLl3Cb7/9hk2bNmn+AJErb2/vQscpFAocOXLEgGmK5/mrAa9btw4nTpzA6tWrkZCQgKCgIOzZs8egeeT9TScTpqamWLBgAW7fvo3Lly8DeHqN/apVqxo5mTTDhg1DcHAwmjdvnq/wRowYYdxwJHvLly/HmDFj0KdPH63h27Ztw/Lly7U2ccrR0aNHjR3hpVlYWGh+Pn/+vOZIRwcHB95ASe6qVq1aYkrieSW98Mi4/vnnnwJLoVevXli9erUREr1dEhMTUa5cOZw9exaffvqpZnhWVpbBs7Aw3iIltfDIuAo7GkehUBjlSJ23SWBgILp27QozMzM0aNBAs7/l4sWLRtl/xMIgIp0UCoXWrUFfHEf64+PjA5VKhQcPHmgdFeXo6GiU+/BwpzcR6SS324SS8bAwiIhIEm6SIiJJcnJycPv2bQDAu+++y/0XbyEWBhHpJIRAaGgovv/+e5iamkKhUCA3NxeBgYEYNmyYseORAbEwiEinlStX4vfff8fOnTs1J+ndvHkT06ZNg0KhwNChQ42ckAyF+zCISKf27dtj+/btKFOmjNbwhw8fonfv3oiIiDBSMjI03tObiHQyMTHJVxYAUK5cOdlfFodeLxYGEemkUCiQmJiYb3hCQoIR0pAxsTCISKcBAwZg8ODBOHXqFNLT05Geno7ffvsNQ4YM4R0b3zLch0FERdq9ezeWLl2Ku3fvQqFQoHLlyhgxYgS6du1q7GhkQCwMIpLs8ePHAABra2sjJyFj4CYpIpLM2tpaUxZXr14t8Xfjo+JhYRCRTtHR0RgyZAg6deqE0NBQPHz4EJ999hkCAgJQu3ZtY8cjA+ImKSLSqX///vjggw/g6emJI0eO4I8//kCtWrUwdepU2NnZGTseGRALg4h06tSpE8LDwwEAarUaTZs2xcmTJ3kOxluIm6SISKfni0GpVMLBwYFl8ZbiGgYR6eTq6qq505sQAjdv3tQ8BoAdO3YYKxoZGP9MICKdVq5ciZSUFMTGxqJq1aoFXiaE3g4sDCLS6cGDB/jyyy9hbW2N7OxsLFmyBI0bNzZ2LDICFgYR6bRixQps2bIFLi4uOH36NJYtW8bCeEtxpzcR6aRUKuHi4gIAaNSoEdLS0oyciIyFaxhEpFNOTg5u3ryJZ8fHZGdnaz1+fgc4vdl4lBQR6eTt7V3oOIVCgSNHjhgwDRkTC4OIiCThPgwiIpKEhUFERJKwMOit5ubmhjt37hg7BlGJwMKgEsfb2xv169eHm5sbmjRpguDgYM2NfXTp378/tm/frjXswoULqFKlyitnCg4OxrfffqtzGiEE1q9fD19fX7i6uqJFixb49NNPce3atSLnf/fuXTg7OyM3N/eVsxK9LBYGlUihoaG4cOECdu3ahaioKKxYscLYkYo0c+ZMrF+/HhMnTsTZs2cRERGBNm3a4MSJE8aOphNLip5hYVCJVrFiRTRv3hzXr1/Hw4cPERQUhEaNGsHDwwNBQUFISEgAAHz77beIjIxESEgI3NzcEBISAgBwdnbG7du3ATw9v2Du3Lnw8vJCkyZNMGXKFGRmZgIAzpw5gxYtWmD16tVo3LgxmjVrhp07dwIAtm7divDwcPz4449wc3PD0KFD8+WMiYlBWFgYFi5ciMaNG8Pc3BxWVlbo3LkzAgMDAQDHjx9H165d4e7ujpYtW2LJkiWa5wcEBAAAPDw84ObmhgsXLgB4euE/Hx8feHh44JNPPkFsbKzmOb/99hvatWuHBg0aYNq0aQgICNCsYanVaixfvhytWrVC48aN8cUXX2hOyHu2NrN9+3Z4eXnho48+QmBgIDZs2KD1mjp16oRDhw69yj8flTSCqIRp1aqVOHXqlBBCiLi4ONGhQwfx7bffiuTkZPHzzz+LJ0+eiLS0NDFq1CgxbNgwzfMCAgLEtm3btOZVs2ZNERMTI4QQYubMmSIoKEikpKSItLQ0ERQUJBYsWCCEEOL06dPCxcVFfPfddyI7O1scP35c1K9fX6SmpgohhJgwYYJYuHBhoZk3bdokvLy8dL6u06dPi6tXr4q8vDxx5coV0bhxY3Ho0CEhhBB37twRNWvWFDk5OZrpDx06JNq0aSNu3LghcnJyxLJly4Sfn58QQoikpCTh5uYmIiIiRE5Ojli7dq2oXbu25vVv375dtGnTRvz3338iPT1djBgxQowbN05rWePHjxePHz8WGRkZYv/+/aJnz56aZV+5ckV4enqKrKwsna+J3ixcw6ASacSIEVCpVPD394eHhweGDh2K8uXLo127drCyskLp0qUxbNgwnDt3TtL8hBDYtm0bvvrqK9jY2KB06dIICgrC/v37NdOYmppixIgRMDMzQ8uWLVGqVClER0dLmn9qairs7e11TtOwYUM4OztDqVSiVq1a6NixI86ePVvo9Fu2bEFgYCCqV68OU1NTDB06FFeuXEFsbCx+/fVX1KhRA23btoWpqSkGDBiAd955R/Pc8PBwfPzxx6hSpQqsra0xZswYHDhwQGvz06hRo1CqVClYWlqidevWiImJQUxMDABgz5498PHxgbm5uaTXT28GXhqESqRly5ahSZMmWsMyMjIwe/ZsnDx5Eg8fPgQAPH78GHl5eTAxMdE5v+TkZGRkZKB79+6aYUIIqNVqzWMbGxutGwdZWVnhyZMnkvLa2Njg/v37Oqf566+/sGDBAly/fh05OTnIzs5G+/btC50+Li4Os2bNwty5c7UyJyYm4t69e3BwcNAMVygUWo/v3buHSpUqaR5XqlQJubm5SEpK0gx7fnoLCwv4+Phg7969GDlyJPbt24fFixdLeu305mBh0Btj9erViI6OxrZt22Bvb48rV66ga9eummse6VK+fHlYWlpi//79qFixYrGXrVAodI5v3LgxQkJCcOnSJdSrV6/AacaOHYuAgACsWrUKFhYWmDlzJlJSUgqdv6OjI4YOHYrOnTvnG3f79m0kJiZqHgshNPtzAKBChQpa+zvi4uJgamoKOzs7zXQvLrNbt2744osv0KBBA1hZWcHNzU3na6Y3DzdJ0Rvj8ePHsLCwQNmyZZGamoqlS5dqjX/nnXcKPedCqVSiV69emDVrluav7MTERJw8eVLSsu3s7HD37t1Cx1erVg3+/v4YO3Yszpw5g+zsbGRlZWH//v1YuXKlJn+5cuVgYWGBv//+G/v27dM839bWFkqlUit/nz59sHLlSly/fh0AkJaWhoMHDwIAWrZsiWvXruHw4cPIzc1FWFgYHjx4oHmur68v1q1bhzt37uDx48f49ttv4ePjo/PWq25ublAqlZgzZ06BJUVvPhYGvTE++ugjZGVloVGjRvDz80Pz5s21xg8YMAARERHw8PDA119/ne/5CHPMOAAAAWZJREFU48ePR9WqVdG7d2+4u7vj448/lryPomfPnrhx4wZUKhWGDx9e4DSTJk1Cv379EBISAg8PD7Rp0waHDh1Cq1atAABTp07F4sWL4ebmhmXLlsHHx0fzXCsrKwwdOhR9+/aFSqXCxYsX8eGHH2Lw4MEYM2YM3N3d4evri19//RXA04JZtGgR5s+fj4YNG+LGjRuoW7cuzMzMAAA9evRA586dERAQgNatW8Pc3ByTJ08u8nV26dIF//77L7p06SLpfaE3Cy8+SPQWUKvVaNGiBRYsWIBGjRq99Hx2796NrVu3YvPmza8xHZUUXMMgekOdPHkSjx49QnZ2NkJDQwEArq6uLz2/jIwMbNq0CX5+fq8rIpUwLAyiN9SzzVYNGzbEsWPHsGzZMlhaWr7UvE6ePInGjRvDzs4Ovr6+rzkplRTcJEVERJJwDYOIiCRhYRARkSQsDCIikoSFQUREkrAwiIhIkv8HIzJyc9rM5rUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gvw9Eh1M_-_d"
      },
      "source": [
        "\n",
        "\n",
        "Above bar graph represents the category of participating patients. It can be infered that most of the participants fall under Healthy Control(HC) and Parkinsons Disease (PD) category. Thus we will consider records from this two categories for training model. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "79voWet4HrX5",
        "outputId": "1a23850b-ae4e-46e7-e5e7-8665451c6998",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "merged_data[\"info\"].GENDER.replace([0, 1, 2], ['F bearing child', 'F Normal', 'M']).reset_index().groupby(\"GENDER\").size().plot(kind='bar', title=\"Gender of selected participants\", rot=0, color = 'c');"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nY40oece_FhS"
      },
      "source": [
        "Above bar graph represent distribution of participants on basis of gender. Also, there is a different category marked fro female participant who are pregnent. It can be depicted that their are higher number of male participants in this data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a0ATr5j2IWM_",
        "outputId": "fd3b3938-8a13-4137-e9a2-b65a45a95b2c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "hc_birthdt = merged_data[\"info\"].BIRTHDT.reset_index()[merged_data[\"info\"].reset_index()['ENROLL_CAT'] == 'HC'].rename(columns={'BIRTHDT':'HC'})\n",
        "pd_birthdt = merged_data[\"info\"].BIRTHDT.reset_index()[merged_data[\"info\"].reset_index()['ENROLL_CAT'] == 'PD'].rename(columns={'BIRTHDT':'PD'})\n",
        "concat_birthdt = pd.concat([pd_birthdt.PD, hc_birthdt.HC], axis=1)\n",
        "plot_3 = concat_birthdt.plot(kind='hist', title=\"Histogram of age\", alpha=0.7, bins=15, ax=plt.gca(), color = ['r','y'])\n",
        "plot_3.set_xlabel(\"Birthdate\");"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-JWZ6lF5oIew"
      },
      "source": [
        "Above Histogram represents the frequency distrubution of participants on basis of their age groups for two dominating categories HC and PD. It can be infered that most of the participating patients suffereing from PD fall under age group above 40 years."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "63YFyEEAt7oA"
      },
      "source": [
        "###Feature Engineering"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qd6B5jVyajeu"
      },
      "source": [
        "**STEP 1: Time Series Vectorisation & Imputation**\n",
        "\\\n",
        "\\\n",
        "Below steps are perfomed to Vectorize time series data into one series, that is patients data from each visit and handle missing values using ***interpolate()*** function. Vectorization technique eliminates looping hops, generating better performing model and reduces the computation load by 20 – 30%. In this technique data points of each covariate from every visit of patient is pivoted against its unique patient ID, creating a single row of record for each patient ID."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R1_PuriqSAkE"
      },
      "source": [
        "d1 = merged_data['updrs1'].drop('INFODT',axis=1) #drop unwated column\n",
        "d1 = d1.set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO') #set index and sort data \n",
        "d1 = d1.unstack().reset_index().set_index('PATNO') #pivot sorted data and reset index\n",
        "d1 = d1.interpolate(method='linear', axis=1) #handeling missing values using linear function\n",
        "\n",
        "d2 = merged_data['updrs1pq'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d2 = d2.unstack().reset_index().set_index('PATNO')\n",
        "d2 = d2.interpolate(method='linear', axis=1)\n",
        "\n",
        "d3 = merged_data['updrs2pq'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d3 = d3.unstack().reset_index().set_index('PATNO')\n",
        "d3 = d3.interpolate(method='linear', axis=1)\n",
        "\n",
        "d4 = merged_data['updrs3'].drop(['PAG_NAME','EXAMTM','PD_MED_USE','ON_OFF_DOSE','ANNUAL_TIME_BTW_DOSE_NUPDRS'],axis=1)\n",
        "d4 = d4.set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d4 = d4.unstack().reset_index().set_index('PATNO')\n",
        "d4 = d4.interpolate(method='linear', axis=1)\n",
        "\n",
        "d5 = merged_data['moca'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d5 = d5.unstack().reset_index().set_index('PATNO')\n",
        "d5 = d5.interpolate(method='linear', axis=1)\n",
        "\n",
        "d6 = merged_data['stai'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d6 = d6.unstack().reset_index().set_index('PATNO')\n",
        "d6 = d6.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d7 = merged_data['gds'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d7 = d7.unstack().reset_index().set_index('PATNO')\n",
        "d7 = d7.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d8 = merged_data['quip'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d8 = d8.unstack().reset_index().set_index('PATNO')\n",
        "d8 = d8.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d9 = merged_data['scopa_aut'].drop(['SCAU23AT','SCAU26AT','SCAU26BT','SCAU26CT','SCAU26DT'],axis=1)\n",
        "d9 = d9.set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d9 = d9.unstack().reset_index().set_index('PATNO')\n",
        "d9 = d9.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d10 = merged_data['sft'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d10 = d10.unstack().reset_index().set_index('PATNO')\n",
        "d10 = d10.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d11 = merged_data['rem_rbd'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d11 = d11.unstack().reset_index().set_index('PATNO')\n",
        "d11 = d11.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d12 = merged_data['epworth'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d12 = d12.unstack().reset_index().set_index('PATNO')\n",
        "d12 = d12.interpolate(method='linear', axis=1)\n",
        "\n",
        "d13 = merged_data['hvlt'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d13 = d13.unstack().reset_index().set_index('PATNO')\n",
        "d13 = d13.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d14 = merged_data['lns'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d14 = d14.unstack().reset_index().set_index('PATNO')\n",
        "d14 = d14.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d15 = merged_data['neuro_cranial'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d15 = d15.unstack().reset_index().set_index('PATNO')\n",
        "d15 = d15.interpolate(method='linear', axis=1)\n",
        "\n",
        "d16 = merged_data['sdm'].set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d16 = d16.unstack().reset_index().set_index('PATNO')\n",
        "d16 = d16.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "d17 = merged_data['benton'].drop_duplicates(['PATNO','EVENT_ID'], keep='first').set_index(['PATNO','EVENT_ID']).sort_index(level='PATNO')\n",
        "d17 = d17.unstack().reset_index().set_index('PATNO')\n",
        "d17 = d17.interpolate(method='linear', axis=1)\n",
        "\n",
        "arr_of_vect_datasets = [d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16,d17]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3NVhDzGFjmYj"
      },
      "source": [
        "**STEP 2: Normalization\\Standardization**\n",
        "\n",
        "\\\n",
        "Normalization\\Standardization techniques are adopted to uniformly scale data points. As different covariates in dataset are measured on different scales and thus do not contribute equally during training of machine learning model, which might end up creating a bias.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FpuV6heOxv3C"
      },
      "source": [
        "**Min-Max normalization**  technique is performed to scale value of data points between 0 and 1. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UPGbIPz8zCUa",
        "outputId": "65e90c01-503d-4bcf-cb4f-51e747c1c1bf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 256
        }
      },
      "source": [
        "minmax = {}\n",
        "i=1\n",
        "for d in arr_of_vect_datasets:\n",
        "    dataset = 'd' + str(i)\n",
        "    i = i+1\n",
        "    dataset_columns = eval(dataset).columns.levels[0][0:-1]\n",
        "    minmax[dataset] = pd.DataFrame(index=eval(dataset).index, columns=eval(dataset).columns)\n",
        "    for col in dataset_columns:\n",
        "        minmax[dataset][col] = (eval(dataset)[col]-eval(dataset)[col].min().min())/(eval(dataset)[col].max().max()-eval(dataset)[col].min().min() )\n",
        "\n",
        "db_minmax = pd.concat([minmax['d1'], minmax['d2'], minmax['d3'], minmax['d4'], minmax['d5'], minmax['d6'] , minmax['d7'], \n",
        "                  minmax['d8'], minmax['d9'], minmax['d10'], minmax['d11'], minmax['d12'], minmax['d13'], minmax['d14'],\n",
        "                  minmax['d15'], minmax['d16'],minmax['d17']], axis=1)\n",
        "db_minmax = db_minmax.interpolate(method='linear', axis=1, limit=10, limit_direction='both')\n",
        "\n",
        "db_minmax.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead tr th {\n",
              "        text-align: left;\n",
              "    }\n",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"7\" halign=\"left\">NP1COG</th>\n",
              "      <th colspan=\"7\" halign=\"left\">NP1HALL</th>\n",
              "      <th colspan=\"7\" halign=\"left\">NP1DPRS</th>\n",
              "      <th colspan=\"7\" halign=\"left\">NP1ANXS</th>\n",
              "      <th colspan=\"7\" halign=\"left\">NP1APAT</th>\n",
              "      <th colspan=\"5\" halign=\"left\">NP1DDS</th>\n",
              "      <th>...</th>\n",
              "      <th colspan=\"4\" halign=\"left\">CN7RSP</th>\n",
              "      <th colspan=\"6\" halign=\"left\">CN8RSP</th>\n",
              "      <th colspan=\"6\" halign=\"left\">CN910RSP</th>\n",
              "      <th colspan=\"6\" halign=\"left\">CN11RSP</th>\n",
              "      <th colspan=\"6\" halign=\"left\">CN12RSP</th>\n",
              "      <th colspan=\"6\" halign=\"left\">SDMTOTAL</th>\n",
              "      <th colspan=\"6\" halign=\"left\">JLO_TOTRAW</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>EVENT_ID</th>\n",
              "      <th>BL</th>\n",
              "      <th>V02</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V02</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V02</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V02</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V02</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V02</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>...</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "      <th>BL</th>\n",
              "      <th>V04</th>\n",
              "      <th>V06</th>\n",
              "      <th>V08</th>\n",
              "      <th>V10</th>\n",
              "      <th>V12</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PATNO</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>3000</th>\n",
              "      <td>0.25</td>\n",
              "      <td>0.125</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.125</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.519608</td>\n",
              "      <td>0.539216</td>\n",
              "      <td>0.500000</td>\n",
              "      <td>0.519608</td>\n",
              "      <td>0.500000</td>\n",
              "      <td>0.519608</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.933333</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.933333</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3001</th>\n",
              "      <td>0.00</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.333333</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.411765</td>\n",
              "      <td>0.352941</td>\n",
              "      <td>0.411765</td>\n",
              "      <td>0.470588</td>\n",
              "      <td>0.470588</td>\n",
              "      <td>0.441176</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.8</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.933333</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3002</th>\n",
              "      <td>0.25</td>\n",
              "      <td>0.250</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.75</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.125</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.333333</td>\n",
              "      <td>0.166667</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.666667</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.401961</td>\n",
              "      <td>0.372549</td>\n",
              "      <td>0.294118</td>\n",
              "      <td>0.431373</td>\n",
              "      <td>0.421569</td>\n",
              "      <td>0.372549</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.8</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.733333</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.866667</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3 rows × 1595 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "         NP1COG                     ... JLO_TOTRAW                              \n",
              "EVENT_ID     BL    V02   V04   V06  ...        V06       V08       V10       V12\n",
              "PATNO                               ...                                         \n",
              "3000       0.25  0.125  0.00  0.25  ...   0.933333  0.866667  0.933333  1.000000\n",
              "3001       0.00  0.000  0.25  0.00  ...   1.000000  0.866667  0.933333  1.000000\n",
              "3002       0.25  0.250  0.25  0.50  ...   0.866667  0.733333  1.000000  0.866667\n",
              "\n",
              "[3 rows x 1595 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tsT5wwzbuxr1"
      },
      "source": [
        "###Dimensionality Reduction\n",
        "\n",
        "Dimensionality reduction techniques are adpoted to scale down the number of covariate, dimensionality reduction technique is applied to group corelated features into new component vectors and prevent loss of information. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G2GIoeMF4PcB",
        "outputId": "a4f2ce11-3380-45dc-ea66-30c7311d7f1c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        }
      },
      "source": [
        "db_selected = db_minmax #db_zs\n",
        "db_selected_labled = pd.concat([db_selected, merged_data[\"info\"].ENROLL_CAT], axis=1) # labeling of selected subjects\n",
        "db_selected_labled.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>(NP1COG, BL)</th>\n",
              "      <th>(NP1COG, V02)</th>\n",
              "      <th>(NP1COG, V04)</th>\n",
              "      <th>(NP1COG, V06)</th>\n",
              "      <th>(NP1COG, V08)</th>\n",
              "      <th>(NP1COG, V10)</th>\n",
              "      <th>(NP1COG, V12)</th>\n",
              "      <th>(NP1HALL, BL)</th>\n",
              "      <th>(NP1HALL, V02)</th>\n",
              "      <th>(NP1HALL, V04)</th>\n",
              "      <th>(NP1HALL, V06)</th>\n",
              "      <th>(NP1HALL, V08)</th>\n",
              "      <th>(NP1HALL, V10)</th>\n",
              "      <th>(NP1HALL, V12)</th>\n",
              "      <th>(NP1DPRS, BL)</th>\n",
              "      <th>(NP1DPRS, V02)</th>\n",
              "      <th>(NP1DPRS, V04)</th>\n",
              "      <th>(NP1DPRS, V06)</th>\n",
              "      <th>(NP1DPRS, V08)</th>\n",
              "      <th>(NP1DPRS, V10)</th>\n",
              "      <th>(NP1DPRS, V12)</th>\n",
              "      <th>(NP1ANXS, BL)</th>\n",
              "      <th>(NP1ANXS, V02)</th>\n",
              "      <th>(NP1ANXS, V04)</th>\n",
              "      <th>(NP1ANXS, V06)</th>\n",
              "      <th>(NP1ANXS, V08)</th>\n",
              "      <th>(NP1ANXS, V10)</th>\n",
              "      <th>(NP1ANXS, V12)</th>\n",
              "      <th>(NP1APAT, BL)</th>\n",
              "      <th>(NP1APAT, V02)</th>\n",
              "      <th>(NP1APAT, V04)</th>\n",
              "      <th>(NP1APAT, V06)</th>\n",
              "      <th>(NP1APAT, V08)</th>\n",
              "      <th>(NP1APAT, V10)</th>\n",
              "      <th>(NP1APAT, V12)</th>\n",
              "      <th>(NP1DDS, BL)</th>\n",
              "      <th>(NP1DDS, V02)</th>\n",
              "      <th>(NP1DDS, V04)</th>\n",
              "      <th>(NP1DDS, V06)</th>\n",
              "      <th>(NP1DDS, V08)</th>\n",
              "      <th>...</th>\n",
              "      <th>(CN7RSP, V08)</th>\n",
              "      <th>(CN7RSP, V10)</th>\n",
              "      <th>(CN7RSP, V12)</th>\n",
              "      <th>(CN8RSP, BL)</th>\n",
              "      <th>(CN8RSP, V04)</th>\n",
              "      <th>(CN8RSP, V06)</th>\n",
              "      <th>(CN8RSP, V08)</th>\n",
              "      <th>(CN8RSP, V10)</th>\n",
              "      <th>(CN8RSP, V12)</th>\n",
              "      <th>(CN910RSP, BL)</th>\n",
              "      <th>(CN910RSP, V04)</th>\n",
              "      <th>(CN910RSP, V06)</th>\n",
              "      <th>(CN910RSP, V08)</th>\n",
              "      <th>(CN910RSP, V10)</th>\n",
              "      <th>(CN910RSP, V12)</th>\n",
              "      <th>(CN11RSP, BL)</th>\n",
              "      <th>(CN11RSP, V04)</th>\n",
              "      <th>(CN11RSP, V06)</th>\n",
              "      <th>(CN11RSP, V08)</th>\n",
              "      <th>(CN11RSP, V10)</th>\n",
              "      <th>(CN11RSP, V12)</th>\n",
              "      <th>(CN12RSP, BL)</th>\n",
              "      <th>(CN12RSP, V04)</th>\n",
              "      <th>(CN12RSP, V06)</th>\n",
              "      <th>(CN12RSP, V08)</th>\n",
              "      <th>(CN12RSP, V10)</th>\n",
              "      <th>(CN12RSP, V12)</th>\n",
              "      <th>(SDMTOTAL, BL)</th>\n",
              "      <th>(SDMTOTAL, V04)</th>\n",
              "      <th>(SDMTOTAL, V06)</th>\n",
              "      <th>(SDMTOTAL, V08)</th>\n",
              "      <th>(SDMTOTAL, V10)</th>\n",
              "      <th>(SDMTOTAL, V12)</th>\n",
              "      <th>(JLO_TOTRAW, BL)</th>\n",
              "      <th>(JLO_TOTRAW, V04)</th>\n",
              "      <th>(JLO_TOTRAW, V06)</th>\n",
              "      <th>(JLO_TOTRAW, V08)</th>\n",
              "      <th>(JLO_TOTRAW, V10)</th>\n",
              "      <th>(JLO_TOTRAW, V12)</th>\n",
              "      <th>ENROLL_CAT</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PATNO</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>3000</th>\n",
              "      <td>0.25</td>\n",
              "      <td>0.125</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.125</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.519608</td>\n",
              "      <td>0.539216</td>\n",
              "      <td>0.500000</td>\n",
              "      <td>0.519608</td>\n",
              "      <td>0.500000</td>\n",
              "      <td>0.519608</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.933333</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.933333</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>HC</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3001</th>\n",
              "      <td>0.00</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.333333</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.411765</td>\n",
              "      <td>0.352941</td>\n",
              "      <td>0.411765</td>\n",
              "      <td>0.470588</td>\n",
              "      <td>0.470588</td>\n",
              "      <td>0.441176</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.8</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.933333</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>PD</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3002</th>\n",
              "      <td>0.25</td>\n",
              "      <td>0.250</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.75</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.125</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.333333</td>\n",
              "      <td>0.166667</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.666667</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.401961</td>\n",
              "      <td>0.372549</td>\n",
              "      <td>0.294118</td>\n",
              "      <td>0.431373</td>\n",
              "      <td>0.421569</td>\n",
              "      <td>0.372549</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.8</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>0.733333</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.866667</td>\n",
              "      <td>PD</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3 rows × 1596 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       (NP1COG, BL)  (NP1COG, V02)  ...  (JLO_TOTRAW, V12)  ENROLL_CAT\n",
              "PATNO                               ...                               \n",
              "3000           0.25          0.125  ...           1.000000          HC\n",
              "3001           0.00          0.000  ...           1.000000          PD\n",
              "3002           0.25          0.250  ...           0.866667          PD\n",
              "\n",
              "[3 rows x 1596 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EwXmWFXKzB90"
      },
      "source": [
        "**Technique 1:** Principal Component Analysis (PCA)\\\n",
        "\\\n",
        "PCA technique is applied to extract the linear combinations of different variables that correlates to the direction with maximum sample variance. Succeeding PCA vector branches out to \u001cfind a direction that gives highest variance such that it is orthogonal (uncorrelated) to preceding ones.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JyYH-cnTICcW",
        "outputId": "4797caa4-880f-4db2-ae52-f9a6ccabda9d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        }
      },
      "source": [
        "pca_targets = ['p1','p2','p3']\n",
        "\n",
        "m_pca = pd.DataFrame(index=db_selected.index, columns=pca_targets) \n",
        "model_pca = sklearnPCA(n_components=3)\n",
        "m_pca[['p1','p2','p3']] = model_pca.fit_transform(db_selected) \n",
        "\n",
        "m_pca.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>p1</th>\n",
              "      <th>p2</th>\n",
              "      <th>p3</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PATNO</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>3000</th>\n",
              "      <td>-2.858799</td>\n",
              "      <td>1.612146</td>\n",
              "      <td>-0.913865</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3001</th>\n",
              "      <td>0.715528</td>\n",
              "      <td>-1.962380</td>\n",
              "      <td>1.095407</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3002</th>\n",
              "      <td>5.141508</td>\n",
              "      <td>-0.210450</td>\n",
              "      <td>1.238283</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             p1        p2        p3\n",
              "PATNO                              \n",
              "3000  -2.858799  1.612146 -0.913865\n",
              "3001   0.715528 -1.962380  1.095407\n",
              "3002   5.141508 -0.210450  1.238283"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m8-aX_Execva",
        "outputId": "cbab7d25-1735-4f3c-a889-e4ff7e715615",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 308
        }
      },
      "source": [
        "fig=plt.figure(figsize=(6,4))  \n",
        "ax=fig.add_subplot(1,1,1)  \n",
        "ax.set_xlabel('Component 1',fontsize = 15)  \n",
        "ax.set_ylabel('Component 2',fontsize = 15)  \n",
        "ax.set_title('Dimensionality reduction using PCA',fontsize=20) \n",
        "colors = ['r','y']\n",
        "targets = ['HC','PD']\n",
        "target = merged_data[\"info\"].ENROLL_CAT\n",
        "#target = db_selected.iloc[:,-1]\n",
        "for target,color in zip(targets,colors):    \n",
        "     indicesToKeep = merged_data[\"info\"].ENROLL_CAT == target  \n",
        "     ax.scatter(m_pca.loc[indicesToKeep,'p1'], m_pca.loc[indicesToKeep,'p2'], c=color, s=40)\n",
        "ax.legend(targets)  \n",
        "ax.grid()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NzDjaYIWypkK"
      },
      "source": [
        "**Technique 2:** Nonnegative Matrix Factorization (NMF)\\\n",
        "\\\n",
        "NMF is an alternative technique for decomposition of variables that assumes data and the components to be non-negative. Unlike PCA, which tends to group both positively as well as negatively correlated components together. NMF tends to finds patterns amoung variable with the same direction of correlation."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qcf2YcARmqNT",
        "outputId": "9b03012f-7d47-41ee-f9de-fb47c7ad7646",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        }
      },
      "source": [
        "nmf_targets = ['n1','n2']\n",
        "#db = db_minmax.drop(['ENROLL_CAT'], axis=1)\n",
        "\n",
        "m_nmf = pd.DataFrame(index=db_selected.index, columns=nmf_targets) \n",
        "model_NMF = decomposition.NMF(n_components=2, init='nndsvda', max_iter=200)\n",
        "m_nmf[['n1', 'n2']] = model_NMF.fit_transform(db_selected)\n",
        "#m_nmf = pd.concat([m_nmf,db_minmax[['ENROLL_CAT']]],axis=1)\n",
        "m_nmf.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>n1</th>\n",
              "      <th>n2</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PATNO</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>3000</th>\n",
              "      <td>0.854110</td>\n",
              "      <td>0.091709</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3001</th>\n",
              "      <td>0.658744</td>\n",
              "      <td>0.498369</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3002</th>\n",
              "      <td>0.325762</td>\n",
              "      <td>0.839209</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             n1        n2\n",
              "PATNO                    \n",
              "3000   0.854110  0.091709\n",
              "3001   0.658744  0.498369\n",
              "3002   0.325762  0.839209"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Tq6FYWMxwQs",
        "outputId": "f4f9f1a3-31b1-44ab-cd2e-923624b36206",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 308
        }
      },
      "source": [
        "fig=plt.figure(figsize=(6,4))  \n",
        "ax=fig.add_subplot(1,1,1)  \n",
        "ax.set_xlabel('Component 1',fontsize = 15)  \n",
        "ax.set_ylabel('Component 2',fontsize = 15)  \n",
        "ax.set_title('Dimensionality reduction using NMF',fontsize=20) \n",
        "colors = ['r','y']\n",
        "targets = ['HC','PD']\n",
        "target = merged_data[\"info\"].ENROLL_CAT\n",
        "#target = m_nmf.iloc[:,-1]\n",
        "for target,color in zip(targets,colors):    \n",
        "     indicesToKeep = merged_data[\"info\"].ENROLL_CAT == target \n",
        "     ax.scatter(m_nmf.loc[indicesToKeep,'n1'], m_nmf.loc[indicesToKeep,'n2'], c=color, s=40)\n",
        "ax.legend(targets)  \n",
        "ax.grid()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MqaN5cT72yZP"
      },
      "source": [
        "**Inferrence**\n",
        "\n",
        "NMF is best suited for clustering with non-negative data and over gaussian data. And as MinMax Normalization technique results in non-negetive datapoint values between 0-1 and data is gaussian in nature, NMF technique is adopeted. It finds\u001c sparse and parts-based representation of the data. Whereas, PCA it is ideal for pattern recognition and dimension reduction."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Bu81WF_Ru6Dm"
      },
      "source": [
        "###**Unsupervised Clustering**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bz1HgDequhjJ"
      },
      "source": [
        "Clustering based on Gaussian Mixture Modelling (GMM) techniques generates data clusters by grouping similar data points based on their feature and correlation between the points. In proposed design data points are grouped into 3 clusters, where each cluster defines the risk of PD progression among the patient. These clusters are marked as Low, Medium and High referencing to progression rate.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vBZsOq0gFReE",
        "outputId": "f234a4dd-cd00-499c-db42-a0ccd4200bf7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        }
      },
      "source": [
        "clus_col = ['GMM']\n",
        "clus_lable = pd.DataFrame(index=db_selected.index, columns=clus_col)\n",
        "clus_lable[ db_selected_labled.ENROLL_CAT == \"HC\" ] = 'HC'\n",
        "\n",
        "gmm_chosen = m_pca #m_pca #m_nmf\n",
        "gmm_chosen_pd = gmm_chosen[ db_selected_labled.ENROLL_CAT == \"PD\" ]\n",
        "gmm_chosen_pd_hc = gmm_chosen[ db_selected_labled.ENROLL_CAT.isin(['HC', 'PD'])]\n",
        "\n",
        "model_gmm = mixture.GaussianMixture(n_components=3, covariance_type='full', random_state=10)\n",
        "model_gmm.fit(gmm_chosen_pd)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GaussianMixture(covariance_type='full', init_params='kmeans', max_iter=100,\n",
              "                means_init=None, n_components=3, n_init=1, precisions_init=None,\n",
              "                random_state=10, reg_covar=1e-06, tol=0.001, verbose=0,\n",
              "                verbose_interval=10, warm_start=False, weights_init=None)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GghiVMqok6Mk",
        "outputId": "3599cc0d-c697-4378-da71-a9489538e678",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 501
        }
      },
      "source": [
        "clus_lable.loc[ db_selected_labled.ENROLL_CAT == \"PD\", 'GMM' ] = model_gmm.predict(gmm_chosen_pd)\n",
        "clus_lable.replace([1,0,2],['PD_l', 'PD_m', 'PD_h'], inplace=True)\n",
        "clus_lable_pd = clus_lable[db_selected_labled.ENROLL_CAT.isin(['PD'])]\n",
        "clus_lable_pd_hc = clus_lable[db_selected_labled.ENROLL_CAT.isin(['HC', 'PD'])]\n",
        "\n",
        "# plot predicted PD and HC\n",
        "plt.figure(figsize=(10, 8))\n",
        "colors = ['r' if i=='HC' else 'g' if i=='PD_l' else 'b' if i=='PD_m' else 'y' if i=='PD_h' else 'black' for i in clus_lable_pd_hc['GMM']]\n",
        "ax = plt.scatter(gmm_chosen_pd_hc.iloc[:,0], gmm_chosen_pd_hc.iloc[:,1], c=colors, alpha=0.8)#gca()\n",
        "p1 = plt.Rectangle((0, 0), 0.1, 0.1, fc='yellow')\n",
        "p2 = plt.Rectangle((0, 0), 0.1, 0.1, fc='blue')\n",
        "p3 = plt.Rectangle((0, 0), 0.1, 0.1, fc='green')\n",
        "p4 = plt.Rectangle((0, 0), 0.1, 0.1, fc='red')\n",
        "plt.legend((p1, p2, p3, p4), ('PD w/ High progression', 'PD w/ Mid progression', 'PD w/ Low progression', 'Healthy Control'), loc='best')\n",
        "plt.title('Clusters based on levels PD pregression')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GfcZg3fbdmUF"
      },
      "source": [
        "Above scatter plot represents the colored distribution of generated clusters.  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1mOeLnHhvLt-"
      },
      "source": [
        "###Ensemble Models\n",
        "\n",
        "For the purpose of study, a comparative study on various classification algorithm is conducted to determine the classification accuracy on this dataset. Models trained for purpose of Study are,\n",
        "1. Nearest Neighbors\n",
        "2. Support Vector Machine\n",
        "3. Decision Tree\n",
        "4. Random Forest Classifier\n",
        "5. AdaBoost\n",
        "6. XGBoost\n",
        "7. Multi-Layer Perceptron Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uWiW106DaYNl"
      },
      "source": [
        "**Step 1: Splitting Dataset into Train and Test data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A9MaeMd7VrFp"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "random.seed(123)\n",
        "#Splitting Data into train and test datasets\n",
        "X = db_selected.loc[clus_lable_pd_hc.index, db_selected.columns.get_level_values(1)=='BL'] \n",
        "X = X.stack().reset_index().set_index('PATNO').drop('EVENT_ID',axis =1)\n",
        "Y = clus_lable_pd_hc.GMM.replace(['HC', 'PD_l', 'PD_m', 'PD_h'], [0, 1, 2, 3])\n",
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bmuAwELFaM9z"
      },
      "source": [
        " **Step 2: Training Classification Models**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1p6Fz-W6Y073"
      },
      "source": [
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.svm import SVC\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier\n",
        "from xgboost import XGBClassifier\n",
        "\n",
        "\n",
        "classifier_names = [\"Nearest Neighbors\", \"SVM\", \"Decision Tree\", \"Random Forest Classifier\", \n",
        "                    \"AdaBoost\", \"XGBoost\"]\n",
        "\n",
        "classifiers = [\n",
        "    KNeighborsClassifier(),\n",
        "    SVC(kernel=\"linear\", C=0.025, class_weight='balanced'),\n",
        "    DecisionTreeClassifier(class_weight='balanced'),\n",
        "    RandomForestClassifier(class_weight='balanced'),\n",
        "    AdaBoostClassifier(),\n",
        "    XGBClassifier()]\n",
        "\n",
        "scores = []\n",
        "for n, c in zip(classifier_names, classifiers):\n",
        "    c.fit(X_train, Y_train)\n",
        "    score = c.score(X_test, Y_test)\n",
        "    scores.append(score)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U2uSLB1AghcR"
      },
      "source": [
        "**Step 3: Model Performance Comparision**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eQokpammf_3k",
        "outputId": "db29a71b-b9a8-422a-efde-a2bb77608271",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "plt.figure(1, figsize=(6, 4))\n",
        "imp, names = zip(*sorted(zip(scores, classifier_names)))\n",
        "plt.barh(range(len(names)), imp, align = 'center')\n",
        "plt.yticks(range(len(names)), names)\n",
        "plt.xlabel('Classifier performance')\n",
        "plt.ylabel('Classifiers')\n",
        "plt.title('Comparision of different classifiers')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U5Uaf1oaaG97"
      },
      "source": [
        "###**Multilayer Perceptron Model**\n",
        "\n",
        "After testing various structural combinations of layers and hyperparametrs, adopted state of art delivers model with best performance. In order to improve performance regularisation using Dropout layers are added and activation fuction Relu and Softmax are used in the structured model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UBBICAvd1zZu",
        "outputId": "99d0115d-03d4-4bbd-fdae-e9858a4f0fd9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        }
      },
      "source": [
        "import keras\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import KFold\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\n",
        "from keras.callbacks import EarlyStopping\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "\n",
        "nn_model  = Sequential()\n",
        "nn_model.add(Dense(units= 100, activation= 'relu', input_dim= 235))\n",
        "nn_model.add(Dropout(0.5))\n",
        "nn_model.add(Dense(units= 10, activation= 'relu'))\n",
        "nn_model.add(Dense(units= 4,  activation= 'softmax'))\n",
        "\n",
        "nn_model.compile(optimizer= 'adam', loss= 'sparse_categorical_crossentropy', metrics= ['sparse_categorical_accuracy'])\n",
        "\n",
        "nn_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_3 (Dense)              (None, 100)               23600     \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 100)               0         \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 10)                1010      \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 4)                 44        \n",
            "=================================================================\n",
            "Total params: 24,654\n",
            "Trainable params: 24,654\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ez-tzzjqJBiB"
      },
      "source": [
        "Early stop function is assigned to training function to monitor the training rate and prevent overfitting. Model checkpoint saves model at every stage of improvement to deliver best perfoming model as resultant."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B3L-frn5C8EA",
        "outputId": "0bcbaff4-1bcf-4c85-a1ed-8294e9c7de33",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "es = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=5)\n",
        "mc = ModelCheckpoint('nn_model.h5', monitor='val_sparse_categorical_accuracy', mode='max', verbose=1, save_best_only=True)\n",
        "history = nn_model.fit(X_train, Y_train, batch_size=32, epochs=50, validation_split=0.25 ,verbose=2, callbacks=[es, mc])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "\n",
            "Epoch 00001: val_sparse_categorical_accuracy improved from -inf to 0.35556, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 1.4158 - sparse_categorical_accuracy: 0.3694 - val_loss: 1.3099 - val_sparse_categorical_accuracy: 0.3556\n",
            "Epoch 2/50\n",
            "\n",
            "Epoch 00002: val_sparse_categorical_accuracy improved from 0.35556 to 0.42222, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 1.2761 - sparse_categorical_accuracy: 0.4104 - val_loss: 1.2423 - val_sparse_categorical_accuracy: 0.4222\n",
            "Epoch 3/50\n",
            "\n",
            "Epoch 00003: val_sparse_categorical_accuracy improved from 0.42222 to 0.45556, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 1.2273 - sparse_categorical_accuracy: 0.4590 - val_loss: 1.1922 - val_sparse_categorical_accuracy: 0.4556\n",
            "Epoch 4/50\n",
            "\n",
            "Epoch 00004: val_sparse_categorical_accuracy improved from 0.45556 to 0.54444, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 1.1312 - sparse_categorical_accuracy: 0.5149 - val_loss: 1.1243 - val_sparse_categorical_accuracy: 0.5444\n",
            "Epoch 5/50\n",
            "\n",
            "Epoch 00005: val_sparse_categorical_accuracy improved from 0.54444 to 0.56667, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 1.0964 - sparse_categorical_accuracy: 0.5634 - val_loss: 1.0727 - val_sparse_categorical_accuracy: 0.5667\n",
            "Epoch 6/50\n",
            "\n",
            "Epoch 00006: val_sparse_categorical_accuracy did not improve from 0.56667\n",
            "9/9 - 0s - loss: 1.0059 - sparse_categorical_accuracy: 0.5933 - val_loss: 1.0321 - val_sparse_categorical_accuracy: 0.5556\n",
            "Epoch 7/50\n",
            "\n",
            "Epoch 00007: val_sparse_categorical_accuracy did not improve from 0.56667\n",
            "9/9 - 0s - loss: 0.9215 - sparse_categorical_accuracy: 0.6343 - val_loss: 0.9891 - val_sparse_categorical_accuracy: 0.5556\n",
            "Epoch 8/50\n",
            "\n",
            "Epoch 00008: val_sparse_categorical_accuracy did not improve from 0.56667\n",
            "9/9 - 0s - loss: 0.8799 - sparse_categorical_accuracy: 0.6642 - val_loss: 0.9541 - val_sparse_categorical_accuracy: 0.5667\n",
            "Epoch 9/50\n",
            "\n",
            "Epoch 00009: val_sparse_categorical_accuracy improved from 0.56667 to 0.63333, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 0.8162 - sparse_categorical_accuracy: 0.7127 - val_loss: 0.8954 - val_sparse_categorical_accuracy: 0.6333\n",
            "Epoch 10/50\n",
            "\n",
            "Epoch 00010: val_sparse_categorical_accuracy did not improve from 0.63333\n",
            "9/9 - 0s - loss: 0.7669 - sparse_categorical_accuracy: 0.7127 - val_loss: 0.8731 - val_sparse_categorical_accuracy: 0.5889\n",
            "Epoch 11/50\n",
            "\n",
            "Epoch 00011: val_sparse_categorical_accuracy improved from 0.63333 to 0.65556, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 0.7060 - sparse_categorical_accuracy: 0.7351 - val_loss: 0.8346 - val_sparse_categorical_accuracy: 0.6556\n",
            "Epoch 12/50\n",
            "\n",
            "Epoch 00012: val_sparse_categorical_accuracy did not improve from 0.65556\n",
            "9/9 - 0s - loss: 0.6436 - sparse_categorical_accuracy: 0.7985 - val_loss: 0.7759 - val_sparse_categorical_accuracy: 0.6556\n",
            "Epoch 13/50\n",
            "\n",
            "Epoch 00013: val_sparse_categorical_accuracy improved from 0.65556 to 0.66667, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 0.7050 - sparse_categorical_accuracy: 0.7164 - val_loss: 0.7561 - val_sparse_categorical_accuracy: 0.6667\n",
            "Epoch 14/50\n",
            "\n",
            "Epoch 00014: val_sparse_categorical_accuracy did not improve from 0.66667\n",
            "9/9 - 0s - loss: 0.5943 - sparse_categorical_accuracy: 0.7985 - val_loss: 0.7710 - val_sparse_categorical_accuracy: 0.6444\n",
            "Epoch 15/50\n",
            "\n",
            "Epoch 00015: val_sparse_categorical_accuracy improved from 0.66667 to 0.67778, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 0.5742 - sparse_categorical_accuracy: 0.7724 - val_loss: 0.7264 - val_sparse_categorical_accuracy: 0.6778\n",
            "Epoch 16/50\n",
            "\n",
            "Epoch 00016: val_sparse_categorical_accuracy did not improve from 0.67778\n",
            "9/9 - 0s - loss: 0.5843 - sparse_categorical_accuracy: 0.7724 - val_loss: 0.7452 - val_sparse_categorical_accuracy: 0.6667\n",
            "Epoch 17/50\n",
            "\n",
            "Epoch 00017: val_sparse_categorical_accuracy did not improve from 0.67778\n",
            "9/9 - 0s - loss: 0.5366 - sparse_categorical_accuracy: 0.8284 - val_loss: 0.7482 - val_sparse_categorical_accuracy: 0.6667\n",
            "Epoch 18/50\n",
            "\n",
            "Epoch 00018: val_sparse_categorical_accuracy improved from 0.67778 to 0.71111, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 0.5326 - sparse_categorical_accuracy: 0.7910 - val_loss: 0.6981 - val_sparse_categorical_accuracy: 0.7111\n",
            "Epoch 19/50\n",
            "\n",
            "Epoch 00019: val_sparse_categorical_accuracy did not improve from 0.71111\n",
            "9/9 - 0s - loss: 0.5132 - sparse_categorical_accuracy: 0.8209 - val_loss: 0.6768 - val_sparse_categorical_accuracy: 0.7111\n",
            "Epoch 20/50\n",
            "\n",
            "Epoch 00020: val_sparse_categorical_accuracy did not improve from 0.71111\n",
            "9/9 - 0s - loss: 0.4574 - sparse_categorical_accuracy: 0.8396 - val_loss: 0.6804 - val_sparse_categorical_accuracy: 0.6889\n",
            "Epoch 21/50\n",
            "\n",
            "Epoch 00021: val_sparse_categorical_accuracy did not improve from 0.71111\n",
            "9/9 - 0s - loss: 0.4322 - sparse_categorical_accuracy: 0.8358 - val_loss: 0.6557 - val_sparse_categorical_accuracy: 0.7111\n",
            "Epoch 22/50\n",
            "\n",
            "Epoch 00022: val_sparse_categorical_accuracy did not improve from 0.71111\n",
            "9/9 - 0s - loss: 0.3601 - sparse_categorical_accuracy: 0.8993 - val_loss: 0.6549 - val_sparse_categorical_accuracy: 0.7111\n",
            "Epoch 23/50\n",
            "\n",
            "Epoch 00023: val_sparse_categorical_accuracy did not improve from 0.71111\n",
            "9/9 - 0s - loss: 0.4095 - sparse_categorical_accuracy: 0.8769 - val_loss: 0.6535 - val_sparse_categorical_accuracy: 0.7000\n",
            "Epoch 24/50\n",
            "\n",
            "Epoch 00024: val_sparse_categorical_accuracy improved from 0.71111 to 0.74444, saving model to nn_model.h5\n",
            "9/9 - 0s - loss: 0.3533 - sparse_categorical_accuracy: 0.9067 - val_loss: 0.6265 - val_sparse_categorical_accuracy: 0.7444\n",
            "Epoch 25/50\n",
            "\n",
            "Epoch 00025: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.3962 - sparse_categorical_accuracy: 0.8582 - val_loss: 0.6290 - val_sparse_categorical_accuracy: 0.7222\n",
            "Epoch 26/50\n",
            "\n",
            "Epoch 00026: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.3501 - sparse_categorical_accuracy: 0.8843 - val_loss: 0.6380 - val_sparse_categorical_accuracy: 0.7111\n",
            "Epoch 27/50\n",
            "\n",
            "Epoch 00027: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.3474 - sparse_categorical_accuracy: 0.9030 - val_loss: 0.6652 - val_sparse_categorical_accuracy: 0.7111\n",
            "Epoch 28/50\n",
            "\n",
            "Epoch 00028: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.3491 - sparse_categorical_accuracy: 0.8843 - val_loss: 0.6390 - val_sparse_categorical_accuracy: 0.7444\n",
            "Epoch 29/50\n",
            "\n",
            "Epoch 00029: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.3157 - sparse_categorical_accuracy: 0.8993 - val_loss: 0.6164 - val_sparse_categorical_accuracy: 0.7333\n",
            "Epoch 30/50\n",
            "\n",
            "Epoch 00030: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.2715 - sparse_categorical_accuracy: 0.8993 - val_loss: 0.6232 - val_sparse_categorical_accuracy: 0.7333\n",
            "Epoch 31/50\n",
            "\n",
            "Epoch 00031: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.3042 - sparse_categorical_accuracy: 0.9030 - val_loss: 0.6207 - val_sparse_categorical_accuracy: 0.7333\n",
            "Epoch 32/50\n",
            "\n",
            "Epoch 00032: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.2883 - sparse_categorical_accuracy: 0.9030 - val_loss: 0.6225 - val_sparse_categorical_accuracy: 0.7333\n",
            "Epoch 33/50\n",
            "\n",
            "Epoch 00033: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.2212 - sparse_categorical_accuracy: 0.9366 - val_loss: 0.6265 - val_sparse_categorical_accuracy: 0.7333\n",
            "Epoch 34/50\n",
            "\n",
            "Epoch 00034: val_sparse_categorical_accuracy did not improve from 0.74444\n",
            "9/9 - 0s - loss: 0.2604 - sparse_categorical_accuracy: 0.8993 - val_loss: 0.6374 - val_sparse_categorical_accuracy: 0.7444\n",
            "Epoch 00034: early stopping\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RZj4tQSlLclB",
        "outputId": "df5890d6-a297-4971-8728-bf4c7dc2e655",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 409
        }
      },
      "source": [
        "plt.figure(figsize=(10,6))\n",
        "plt.plot(history.history['sparse_categorical_accuracy'],linewidth=3.0)\n",
        "plt.plot(history.history['val_sparse_categorical_accuracy'],linewidth=3.0)\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train','validation'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmgAAAGJCAYAAAA+H8/zAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3hUZfrw8e/UTBrpjRJCSSAkhN4RBER6V1FXsaKou25x97XtiuLq6rr+triKu3ZE1FWkhSpVSuhIC5AAgUAIIaT3aef9Y+AkgQCZZJJM4P5cl9flc+bMOU8OyeTOU+5boyiKghBCCCGEcBvapu6AEEIIIYSoTgI0IYQQQgg3IwGaEEIIIYSbkQBNCCGEEMLNSIAmhBBCCOFmJEATQgghhHAzEqAJIZq9F154gb///e+1Onf48OFs27atgXskhBD1IwGaEEIIIYSbkQBNCCHchNVqbeouCCHchARoQohGMXz4cD7++GMmTJhA9+7deemll7h48SKPP/44PXr04OGHH6agoEA9f926dYwbN47evXvz4IMPcuLECfW15ORkpkyZQo8ePfjNb35DRUVFtXtt2LCBSZMm0bt3b+69916OHj1aqz5u3LiRyZMn07NnT4YOHcp7771X7fXdu3dz77330rt3b4YOHcoPP/wAQHl5OW+99RbDhg2jV69e3HfffZSXl7Njxw6GDBly1XO4PMX63nvv8eyzz/L73/+enj17smjRIg4cOMD06dPp3bs3gwcPZs6cOZjNZvX9qampPPLII/Tt25eBAwfy4Ycfkp2dTbdu3cjLy1PPO3z4MP3798disdTqaxdCuBcJ0IQQjWbNmjV89tlnrF69mg0bNjBz5kx+97vfsX37dux2O19++SUAaWlpPPfcc7z00kskJSUxZMgQZs2ahdlsxmw288wzzzBp0iR27tzJ6NGjWbNmjXqP5ORkXnrpJebMmcOOHTuYPn06Tz/9dLUg51o8PT15++232b17N//5z3/4+uuvWbt2LQAZGRnMnDmTBx54gKSkJBYvXkxsbCwAb7/9NocPH+abb75h586d/OEPf0Crrd3H67p16xg9ejS7d+9mwoQJaLVaXnzxRbZv384333xDUlISCxYsAKC4uJhHHnmE2267jc2bN7NmzRoGDBhASEgIffv2ZeXKlep1lyxZwrhx4zAYDLX7xxFCuBUJ0IQQjeaBBx4gODiYsLAwevfuTUJCAl26dMHDw4ORI0eSnJwMwIoVKxg6dCiDBg3CYDDw2GOPUV5ezr59+9i/fz8Wi4WHHnoIg8HA6NGj6dq1q3qPb7/9lunTp9OtWzd0Oh1TpkzBYDDw888/37B//fr1o1OnTmi1Wjp37sy4cePYuXMnAImJiQwcOJDx48djMBgICAggNjYWu93OwoULefnllwkLC0On09GzZ0+MRmOtnkn37t2544470Gq1mEwm4uPj6d69O3q9ntatWzN9+nR27doFOEb4goODefTRR/Hw8MDHx4du3boBMGXKFJYuXQqAzWZj+fLlTJo0qfb/OEIIt6Jv6g4IIW4dwcHB6v97eHhUa5tMJkpLSwG4cOECLVu2VF/TarVERESQlZWFTqcjLCwMjUajvl713HPnzrF48WLmz5+vHrNYLFy4cOGG/du/fz9/+9vfSE1NxWKxYDabGT16NACZmZlERkZe9Z68vDwqKipo06ZNbR7BVcLDw6u109LSeOuttzh06BBlZWXYbDbi4uKu2weAESNGMHv2bM6cOUNaWho+Pj4kJCTUqU9CiKYnI2hCCLcTGhrKuXPn1LaiKGRmZhIWFkZISAhZWVkoiqK+XvXciIgIZs2axe7du9X/9u/fz/jx42943+eee44RI0awadMm9uzZw7333qveJyIigvT09KveExAQgIeHB2fOnLnqNU9PT8rLy9W2zWYjNze32jlVA02AV199lfbt27N69Wr27t3Lb3/722p9qOk+4Ah4x4wZw9KlS1myZImMngnRzEmAJoRwO2PGjGHTpk0kJSVhsVj49NNPMRqN9OjRQ53+mzdvHhaLhTVr1nDw4EH1vXfffTfffPMN+/fvR1EUSktL2bhxI8XFxTe8b0lJCX5+fnh4eHDgwAESExPV1yZMmMC2bdtYsWIFVquVvLw8jhw5glarZdq0afzlL38hKysLm83Gvn37MJvNtGvXjoqKCjZu3IjFYmHu3Lk3XAtXUlKCt7c33t7enDhxgq+//lp97fbbbyc7O5vPP/8cs9lMcXEx+/fvV1+fNGkSixYtYv369RKgCdHMSYAmhHA77du355133uH111+nf//+bNiwgQ8//BCj0YjRaOS9995j0aJF9O3blxUrVjBy5Ej1vV27duX1119nzpw59OnThzvvvFPdbXkjs2fP5l//+hc9evTg/fffZ8yYMeprLVu25KOPPuKzzz6jb9++TJ48Wd0d+vzzzxMTE8Ndd91F3759+dvf/obdbsfX15fZs2fzxz/+kSFDhuDp6XnVlOaVnn/+eRITE+nZsyd/+tOfGDt2rPqaj48Pn376KRs2bGDQoEGMGjWKHTt2qK/36tULrVZLXFwcrVq1qtXXLIRwTxql6jyBEEKIZm3GjBlMmDCBu+++u6m7IoSoBxlBE0KIm8SBAwdITk6uNvInhGieZBenEELcBJ5//nnWrl3Lyy+/jI+PT1N3RwhRTzLFKYQQQgjhZmSKUwghhBDCzUiAJoQQQgjhZiRAE0IIIYRwMzfdJoG8vBLs9oZbVhcU5ENOzo0TXopK8sycJ8/MefLMnCfPzHnyzJwnz6xmWq2GgADva75+0wVodrvSoAHa5XsI58gzc548M+fJM3OePDPnyTNznjwz58kUpxBCCCGEm5EATQghhBDCzUiAJoQQQgjhZm66NWg1sdms5OVlY7Wa632tCxe02O12F/Tq5qXV6vD09MHHxw+NRtPU3RFCCCGanVsiQMvLy8Zk8sLbO7zeAYNer8VqlQDtWhRFwWazUlSUT15eNoGBoU3dJSGEEKLZuSWmOK1WM97eLWQ0pxFoNBr0egP+/kGYzeVN3R0hhBCiWbolAjRAgrNGptFoAdlWLYQQQtTFLROguZNPPvkPFovF6fcdPZrMa6/9sQF6JIQQQgh3IgFaE/jss49qDNCsVut139e5cxdmz/5zQ3VLCCGEEG7iltgk4E7effdtAJ566lE0Gi0RERH4+fmTnn6a0tJSPv98Aa+99kfS009jsZhp1aoNL774Ci1atGDv3t28//4/+eSTL8nMPMfjjz/IxIlT2b59K+Xl5bzwwit069a9ib9CIYQQQtTXLRegrdqRzpKtaVSYbS6/todRx6RB7RjdL/Ka5zz33PMsWvQdc+d+ipeXF2+88SqpqSn8+9//xdPTE4Bf//r3+Pv7A/Df/37AV199wVNP/eqqaxUUFBAfn8CTTz7DmjUr+fDDfzF37qcu/7qEEEKI5uTEuQJS0vMZnBCBr5exqbtTJ7dcgLZ6V3qDBGcAFWYbq3elXzdAq8ntt49QgzOAVasSWbNmFVarhbKyctq0qfl6np5eDBp0GwBxcV3597//UffOCyGEEDeBgydz+Od3B7ArClsOZvLHGb3x9Gh+4c4ttwZtVJ9IPIy6Brm2h1HHqD7OBWcAXl6Vwdn+/ftYvHgh7777HvPmfcvMmU9hNlfU+D6j0aD+v1arxWa7/ho2IYQQ4maWllnIB4sOYVccWQQyc0r5aFmy2m5Oml9IWU+j+0U6PcJVlSsS1Xp5eVNSUoyXl9dVrxUVFeHt7YOfnx9ms5nly5fW615CCCHErSArt5R/fLefCkv1WbKfj19kyeY0pgxp30Q9q5tbLkBzB/fe+wuefXYWHh4mIiIiqr3Wv/9A1qxZyX33TcXPz5/u3XuQnHy4iXoqhBBCuL+CEjPvfvszRaWODAneJj0JHYJJOnwegGXbThEZ5kOvTs2nuo1GUZrhuN915OQUY7dX/5LOnz9NeHhbl1xfSj3V3uXnHhLiS3Z2UVN3p1mRZ+Y8eWbOk2fmPHlmzmvoZ1ZWYeWvC/ZxOstxD6Ney+/v60H7iBb8/bv9HE7LBcDDoOPlGb1oHeLTYH1xhlarISjo2n255dagCSGEEOLmYLXZ+WDRQTU402hg1qR4OrbyQ6vV8OTEOEL8TQBUWGy8t/AAxWXOJ4pvChKgCSGEEKLZsSsKn644wuFTeeqxh0Z3pnt0sNr28TTwq2kJeBgcmwOz88v5z9LD2OzuPxMmAZoQQgghmp3vN55g++EstT15cDuGdGt51XmtQ3x4fHys2j6clsvCjScbpY/1IQGaEEIIIZqVNTvTWbUjXW0P7d6SCYOirnl+r06hTBhY+fqqnelsv7SBwF1JgCaEEEKIZmNHchbfrD+utntEB/PAnTFoNJrrvm/Sbe3o3rFy+vOzlUc5fd59N3xIgCaEEEKIZuHIqVw+TkxW2x1b+fHkxDh02huHM1qNhpkTuhAR5MhBarHaee+HAxSWmBusv/UhAZoQQggh3F56VhHv/XAQ26VUWhFBXjx7VwJGQ+2rA3l66PnVtAS19FNuYQUfLD6E1eZ+mwYkQGsGfvnLJ9i6dTMAH3/8IevWranxvE8++U+t6nGuWLGM9PTTanvLlk28//4/XdNZIYQQwsWy88v4+//2U36plra/j5Hf3dMdH0/DDd55tfBAL56c2IXLE6IpZ/L5Zl2qC3vrGlJJoJl5/PFZ9b7GihXL8PPzJzLSkbx38OChDB48tN7XFUIIIVytqNTM//1vPwWXpiI9PfT87p7uBPmZ6nzNhA7BTB3anoWbHLs51+/NIDLMt8ZdoE1FArRG9vnnH1NYWMCzzz4HQEFBPvffP42XX36NL774BLO5ApvNxowZj3LHHaOuev8bb7xK586xTJs2neLiYt56aw4nT54gMDCIsLAwAgKCANi9eycffTT3qustX76UY8eO8I9//I2PPprLM8/8muzsC2zbtpk///mvAMyf/zmrV68AIDY2jt/85g94eXnxySf/IT39NCUlxZw7l0GrVq15/fW3MZnq/kMihBDOOJ9bigYIC7y6lrG4+VSYbfzz+wNk5ZYCoNdpeHZaV1qH1r8awNj+bTmdVczuoxcAmL/mGC2DvenYyq/e13aFWy5AMx9YScWeJWApd/3FDSY8ek3CmDDmmqeMHj2eJ598iKef/jV6vZ4ff1zFoEFDiI9P4IMPPkan05Gbm8Njjz1I374DaNGixTWv9dlnH+Hl5c2CBQvJz8/n0Ud/wfDhIwGIielc4/XGjZvIypWJ3HffgwwadBvgGFG7LClpK6tXr+DDDz/Fy8ubP/95Np9//jFPP/0sAMeOHeGjj+bh4+PD7373S9asWcnEiVNc8fSEEOK61uw6w7frUtFoNPxyatdqCUnFzcdmtzN3ySFOnisEQAM8MSGOTpEBLrm+RqPhsbGxnM8p5Wx2MVabwvuLDvLKQ30I8PVwyT3qo9ECtLS0NF544QXy8/Px9/fn7bffJioqqto52dnZvPLKK5w9exar1cqsWbOYNGmSS/thPrC6YYIzAEs55gOrrxughYeHExXVge3btzJ48FBWrEjk2Wd/R35+Hn/5yxzOnk1Hp9NTWFhAevpp4uO7XvNa+/bt5je/+QMA/v7+DB06XH2tLtcDx8jbiBF34u3t+Otk4sSp/POff1Nf79u3P76+vgB06RJPRsbZGz8XIYSop1U70vnfBkdqBUVRmLf6KJ0i+6uLvd3dyXOFlJRbiIsKRKu9fjqI5s5ssZFTWE5OQTl5RRV4eXtQVOz8790jp/I4cCJHbd8/MobenV1b7NzDqONX07ry+he7KS6zUFBs5v1FB3n+/h4Y9LXffNAQGu07e/bs2dx///1MmjSJJUuW8MorrzBv3rxq57z11lvEx8czd+5ccnNzmTp1Kn379iUiIsJl/TAmjGrQETRjwtXTklcaO3Y8K1cmEhHRipKSYrp168FvfvM0gwYN4c0330Gj0XDvvVMxmyvq3JV3333Lpde7zGis/KtCq9Vis9nqfU0hhLielTtO892GE9WO5RebWfTTSe4fGdNEvaqdvKIKFqxNYc+xbACiwn15aHRn2ob7NnHP6q7CYiOnoJyLBeXkFJRx8VIwdvHSfw2RtmLcgLaM6NXa5dcFCPH3ZNakOP7v2/3YFYWT5wr5ck0Kj4zpfMPcag2pUQK0nJwckpOT+eyzzwAYP348r7/+Orm5uQQGBqrnHT16lIceegiAwMBAOnfuzMqVK3n00Udd1hdjwpjrjnDdiF6vxWqt33bcoUOH8957/8c338xnzJjxaDQaioqKiIiIQKPRsGvXdjIyztzwOj179mHFimUkJHSnoCCfn37awLBhdwBc93re3t6UlBTXeM3evfsyd+6/uOee+/D09CIxcTF9+vSr19crhLi5nMgo4Ex2MX07h+FlathfIyu2n+b7jZXBWYi/iex8xx/Y6/aeZWDXcKLCr70UpKnY7Qrr9pzlh80nqTBX/iF76nwRc77YxYherZlyW/sGHwEsLbew48gFiuoRNJWbbZeCsDIuFpRTVNq4xcYHxYczdUj7Br1Hl6hA7hneUd3NueVAJm3DfBssKKyNRgnQMjMzCQsLQ6dzDBfqdDpCQ0PJzMysFqDFxcWxYsUKunbtytmzZ9m3bx+tWzfdw2koJpPp0vTmMv73v6UAPPXUL3n33bf55JP/EhvbhQ4dom94nYcffpy//OU17r9/GoGBQXTv3kN97XrXmzhxKv/+999ZsOBLnnnm19WuOWDAIE6cSOXJJx8BoHPnLjz00GOu+LKFEDeBs9nFvL1gL1abwoqk08yaFE/7lg0TICVuO8UPP1XWTOzUxp9f353A+z8c5PCpPBQFvlh1jD/N6O1W04ZpmYXMW3WM01nVs9TrtBpsdgVFgbW7z7LnWDb3jYimV6cQl4/UlJRb+HHXGX7cfZayCqtLr30jWo2GwBYeBPuZCPA10cLXg7KyugV1EUFejOjVulFGskb2bk16VhHbDjlKQH27/jiDu0bgYWyaqU6NoihKQ9/k0KFDPP/88yxfvlw9NnbsWN555x3i4uLUY7m5ubz55pukpKTQsmVLTCYTYWFhvPjii/W6/+HDybRs2bZe1xDOO3fuNHFxXZq6G0IIF3pn/m5+2pehtnVaDQ+P78KkIR1c+kv02x+PMX/VUbWd0DGYPz3aD5OHnnMXi/nlOxuwXJrNmDk5nom3dXDZveuqpMzC/JVHWL4tjaq/WVuH+vD0tG4E+ZmYu/AAP6dmV3tf79gwZk1NcMnO1KJSM0s2nWDZlpOUljdMYKbTaggJ8CQ0wIuwQC9CArwIC3S0QwO8CPIzodM1zzSrFRYbL76/hdQz+XiZ9Hz88kh8vYxN0pdGCdBycnIYNWoUO3bsQKfTYbPZ6NevH2vWrKk2gnalmTNncuedd3L33Xc7ca9i7PbqX9L586cJD3dNgOaKKc5bxeXnHhLiS3a2+9Y7c0fyzJwnz8x5zj6zC3mlvPjf7dT0W6NbhyAeG9+lTolDr7R0SxqLt6Sp7di2ATx7VwIeVTLGL9uaxqLNjnNMRh1vzOzfKDvvanpmiqKw6+gFvl6XSkFx5VSiQa9l/MAoRveNxKDXqufuOJLFN+uOV1urZdRrmTi4HXf2aYO+DsFNUamZNbvOsHbP2WpTquBIzNqrUwjaOgbQer2W4BYmgvxMBPuZ8PfxcGrEsrn9bFZYbOxIziIq3JfIsIZbK6jVaggKuna6kEaZ4gwKCiI2NpbExEQmTZpEYmIisbGxVwVneXl5+Pr6otfrSUpKIiUlhX/961+N0UUhhBA3sHJHuhqcdWzlh81uJy3T8Yt3/4kcZn+6k1mT4ohu7V+n6yuKwpItaSzdeko91iUqgF9Nqx6cAYzu15btyVlk5pRSbraxYG0Kz0y5/i71hnAhv4z5a45x6GRuteNx7QJ58M4YQgOqj4ppNBr6dwknoX0QCzedZOO+DBTAbLXz/cYTJB06z4OjOhHTpnbPsLDUzOqd6azfk0GFpXpgFhHkxYSBUfSNDXOrKWB352HQuUXC2kbbxfnqq6/ywgsv8MEHH9CiRQvefvttwDFK9uyzz9K1a1cOHDjAG2+8gVarJSAggA8//BBPT8/G6qIQQohryCuqYOvBTLU9dUh7Orb24/uNJ1iz64x6zttf7WPKkHaM6d/WqREbRVFYvDmNZdtOqcfi2gXyq6lda6y1aNBrmTGqE28v2AfAnmPZ7D9+kW4dGyc3mtVmZ9WOdJZtO6VOtQL4eRu5745o+nQOve6Ur5fJwIOjOjEwPpx5q49x5oJj41bGxRLe+movtyVEcPewjtcckSwsMbNqRzrr953FbKk+q9My2JsJA6Po0zlUArNmrFGmOBvTtaY4w8IiXbI+QqY4a0dR7GRlnSU8PLLZDW+7A3lmzpNn5jxnntm361NZvdMRiHVo1YKXHuilfqb+nHqRT5YnU1JlzVNcu0Bmju9CC+8br99RFIVFm0+SuK2yRnB8e0dwdqNcVJ8sT2brQcei7qAWJv78eL8GXdQdEuLLlj3pzFt9jMycUvW4BhjWsxVTh3RwemerzW5n7e6zLN6cVm0UzMfTwPThHRkYH64+64LiClbuSGfjvgzMV/wuahXizcRB7eo1ndkQ5GezZjea4rwlArSLFzMxmbzw9m5R7yBNArTrUxQFm81KUZFjh1VgYKj8cNaBPDPnyTNzXm2fWXGZhT98sE0NHp6dlnBVFv/cwnI+XHqY42cL1GN+3kaemBhHbNtrZ35XFIWFm06yYntlcJbQIYhnpsTXKlFoUamZl/67XQ0Ox/SL5O5hHW/4vrooKjWzLCmdtbvSqx2PDPVhxujO9d7NmlNQzoK1KexLvVjteOdIfybf1p49x7LZ+HNGtRE7gNYhPkwcFEVPNwvMLpOfzZpJgAbYbFby8rKxWuufPE+r1WK3S4B2PVqtDk9PH3x8/NBoNPLDWQfyzJwnz8x5tX1mS7akseTSov3WId68+mjfGgMBm93O4s1pLE+qDLY0wIRBUUwc1O6q6TZFUfh+4wlW7qgMeBzBWVd1UX1tbN5/js9WOnZ86rQaZj/cxyW1Gqv6+fhFPl1+hOIq6SI8jDqm3NaeEb1aodO6btfivtRsvvoxhdzC6ycXjwz1YeLgdnSPDnbLwOwy+dmsmVtsEmhqOp2e4GDXVCOQbzQhxK2krMLK2t2Via7HXmdtmU6rZdrQDnRq489HickUlVpQgKVbT5FyJp+ZE+LUnZaKovDdhhOs2lkZnHXvGMxTk+OdCs4ABidEsPVgJilnC7DZFb5YfZQXH+jlsqBl7e4zfL0utdru1V4xIdx3RzSBLUwuuUdVPaJDiG0bwNItp1iz6wz2K8ZR2ob5MnFwFN07BjdppnvRsJpnohIhhBCNYtPP59TpwxB/E31ib1wLMb59EK8+0pfOkZU7EY+m5/PqZzs5dDIHRVH4dv3xasFZj+hgnp7ifHAGjp2RD47ujO7SCN2JjEJ+2n/O6etcyW5X+HptKgvWVgZnwf6ePHtXAs9M7dogwdllJqOee4Z3ZPYjfejY2g+AdhG+PHtXAq883Jse0a5Pbivcyy0xgiaEEMJ5Fqud1VXWW43p37bWU3kBvh78/t4eLNt2iqVb0lCAolIL//e//cS09iOlylq1njEhzJoUV6f8X5e1CvZmdL9IdXr1+w0n6BEdgl8tNinUpMJi479LD1dbD9a+ZQtee2IglnLX15q8ljahPrz0QC9Kyy14meqfY040HzKCJoQQokZbD2WqiVf9fIwMinduqYhWq2HS4Hb8/r4e+PlUBkpVg7NeneofnF02YWAUIf6OUa3SCiv/W59ap+sUlJj564J91YKzXjEh/L/7euDfCMlwayLB2a1HAjQhhBBXsdntrNpeOXo2qk9knaYfwVEJ4LVH+hLXrnpy8t6dQnhyomuCMwCjQceDd3ZS20mHs0g+lXudd1zt3MUS3pi3m7TMQvXYqL5teGpKfI352IRoKBKgCSGEuMquoxe4kF8GgLdJz9Du9cus3sLbyG/v6cY9wzoSEeTFnX3a8IQLg7PL4tsH0bfKOrkvVx/DYrVd5x2Vjp7O480v93CxoBwAjQYeuDOG6cOj3XqXpLg5yRo0IYQQ1SiKwooqqTJG9GqNp0f9f11oNRpG94tkdL/Iel/reu4dEc3BkzmUVdjIyitjedJpJt/W/rrvSTp0nk9XHMF2KU2T0aBl1qR4ujdSZQIhriQjaEIIIarZfyKHs9klgKMu4R292zRxj5zj7+PBtKEd1PaK7afJzCmp8VxFUVi6NY2PEpPV4MzP28gLv+gpwZloUhKgCSGEUCmKwvKkU2p7aPeW16wH6c5u796KdhGOzP5Wm8L8NSlcmZfdarPz6YojLN6cph5rFezNH2f0Jiq8flUBhKgvCdCEEEKoUs7kcyLDsUBer9Mwqm/DTkc2FK1Ww0OjO6lrx46cziPp8Hn19dJyC3//3361jidAl6gAXnygF0F+DZffTIjakgBNCCGEKrHK2rOB8RFq5v/mKDLMlzt6t1bb364/TnGZhYsFZfxl/l6OnM5TXxvcNYLf3N3N6ULnQjQU+U4UQggBwKnzhRxOc6Sl0GhgTP/mOXpW1eTb2rH72AVyCysoKrXwcWIyp88XUVBSmWx2ym3tGD8wSjLzC7ciI2hCCCEAqhU579M5lLAArybsjWuYjHp+cUeM2j5wIkcNzvQ6DTMndGHCoHYSnAm3IwGaEEIIMnNK2HssW22PGxDVdJ1xsR4xIfSIrr4j09uk57np3RkQF95EvRLi+iRAE0IIwYrtp7m8xzGhQxBtQn2atD+u9ouRMWout2A/Ey892ItOkQFN3Cshrk3WoAkhxC0up6Cc7Yez1Pb4m2j07LLAFib+9FBvjp8toGdMsNS2FG5PAjQhhLjFrdqZriZp7dTGn46t/Zq4Rw0jPNCL8MDmv65O3BpkilMIIW5hhSVmftp/Tm2PG9C2CXsjhLhMAjQhhLiF/bj7DBarHYC2Yb7EtQts4h4JIUACNCGEuGWVlFlYv/es2h43oK2kmxDCTUiAJoQQt6gV29Ioq7ABjvVZPWNCmrhHQojLJEATQohm6sri3zekL+MAACAASURBVM6osNhY8tMJtT2mfyRarYyeCeEuZBenEEI0M3lFFXy7PpWfj1/Ew6Aj2M9EkJ8nwS1MBPmZLrVNBLUwqbm/rrTlQCYFxY6M+oEtPCRhqxBuRgI0IYRoJmx2O+v3ZPDD5pNUmB1Tk2aLnaJSC2mZRTW+x8fTQFCLKkHbpQBu1Y7Ksk6j+0ai18mEihDuRAI0IYRoBtIyC5m36hins2oOxK6luMxCcZnlmu/z9TJwW7eWruiiEMKFJEATQgg3VlpuZdFPJ1m/9yxVV5xFBHnx4J2dCAv04mJBGRcLyrlYUE5OQTk5l9o5heVYbddfpzaydxs8DLqG/SKEEE6TAE0IIdyQoijsOnqBr9elqmvFAAx6LRMGRjG6X+W0ZICvB9Gtr76GXVEoLDFfCt7KLgVvjkAut6iC6DYB3NmnTWN9SUIIJ0iAJoS4pZWWWzh3sZT2LVu4zS7GC/llzF9zjEMnc6sdj28XyAN3xhAaULtyRVqNBn8fD/x9POjY6uryTSEhvmRnOzdlKoRoHBKgCSFuWeVmK3+et4fzuaXEtQvk13clNOlieavNzqod6SzbdkrN7g/g523kvjui6dM5VBLJCnGLkABNCHHLWrPrDOdzSwE4nJbL/9Yf5/6RMU3Sl2PpecxbfYzMnFL1mAYY3rM1U4a0x8skH9dC3ErkJ14IcUsqLDWzakd6tWNr95ylTZgPtyU03q7GolIz3204wZaDmdWOR4b58NDozrSLaNFofRFCuA8J0IQQt6TEbacov5RLTAPqDskvVx+jZbA3HVpevWbL1bYezOTb9ccpLrOoxzyMOqbe1p7hvVqh00puMiFuVfLTL4S45WTnl7Fhb4bafnxCF1qHeANgtSm8/8NB8osrGuz+iqKwePNJPll+pFpw1ismhDce78fIPm0kOBPiFiefAEKIW86izSex2R1jZh1b+dG/Sxi/nJaA96V1XvnFZt5fdLDaQn1XcQRnaSzdeko9FtTCxLN3JfDM1K4EtjC5/J5CiOZHAjQhxC0lPauIHYez1PZdt3dAo9EQ6u/JrMnxXN4keSKjkK9+PFavguRXUhSFRZtPsmzbKfVYfPtA/vx4P7p3DHbZfYQQzZ8EaEKIW8r3m06o6826dwwmpo2/+lpcVCD3DOuotn/an8nGfRm4gqIoLNx0ksRtlTUwEzoE8aupXfEwSiZ/IUR1EqAJIW4ZR07lqslfNRqYOrT9Vefc2acNA+LC1PaCtakcS8+r130VReH7jSdYsb16cPbMlK4Y9BKcCSGuJgGaEOKWoCgK3286obYHxUfQOsTnqvM0Gg0Pje5M23BfAGx2hQ8WHyKnoLzO9/1uwwlWVknp0b1j8KXgTD6ChRA1k08HIcQtYc+xbNIyHWWN9Dotk29rd81zjQYdv5raFV8vAwBFpRb+/cNBzBabU/dUFIVv1x9n1c7K4KxHdDBPT4mX4EwIcV3yCSGEuOlZbXYWVhk9u6NX6xvulgxsYeKZKV3RXarPeTqriM9XHa31pgFFUfh6XSprdp1Rj/WMCeGpyfFNWk5KCNE8NNqnRFpaGtOnT2fUqFFMnz6dU6dOXXVOTk4OTzzxBBMmTGDMmDG8+uqrWK3WxuqiEOImtflAJll5ZQB4eugZO6Btrd4X08af+++IVtvbD2exeueZ67zDQVEUFqxNZe3us+qxXp1CmDUpToIzIUStNNonxezZs7n//vtZvXo1999/P6+88spV53z44Yd06NCBZcuWsXTpUg4fPsyaNWsaq4tCiJtQhdnG0i1panvcgLb4eBpq/f7be7RiSLfK0k/fbTzO4bTca56vKApf/ZjCuj2VwVnvzqE8OVGCMyFE7TXKp0VOTg7JycmMHz8egPHjx5OcnExubvUPOY1GQ0lJCXa7HbPZjMViISwsrKZLCiHckNXm+sSu9bVm9xkKSswABPh6cEev1k69X6PR8IuRMXRo5aiJqSjw4ZJDXMgrvepcu6Iwf00K66tUKegbG8qTE7tIcCaEcEqjfGJkZmYSFhaGTufYTq7T6QgNDSUzs3px4Keffpq0tDQGDx6s/terV6/G6KIQN4XFm0/yxDsb+XptaqPe12K189cFe/nl339i6dY0lyZ3rY+iUjMrq6S2mDS4HUaD82ktDHotz0zpir+PEYCScivv/XCQcnPlEgy7ojB/9TE27KsenM2c0EXKNgkhnOZWxdJXrVpFp06d+OKLLygpKWHmzJmsWrWK0aNH1/oaQUFXb5t3tZAQ3wa/x81GnpnznH1mpzIL1fJBP+4+w5Th0bQJa5znvnHPGY6m5wOweHMaBaUWfnVPj0bfqXjlM1u85KBaEL11qA+Th0Wjq+NIVkiIL396rD/P/3sLVpudjOwSvvwxlRdm9AHgg4X72fjzOfX8oT1a89v7etT5fo1FfjadJ8/MefLMnNcoAVpERARZWVnYbDZ0Oh02m40LFy4QERFR7bz58+fz5ptvotVq8fX1Zfjw4ezYscOpAC0npxi7veH+eg8J8SU7u6jBrn8zkmfmvLo8s69WJFdrL998gmlDO7iyW9e0KulUtfaGPWfJzC7mmald8TbVfr1XfVz5zC7ml7Fia+Xas8mD25GbW1KvewR46pkxqhOfrjgCQNLBTD5dcpDcwnI2H6icERgQF86DI6Prfb+GJj+bzpNn5jx5ZjXTajXXHVRqlD/tgoKCiI2NJTExEYDExERiY2MJDAysdl7r1q356aefADCbzSQlJREdHX3V9YQQ1V3IK2XHkaxqx5IOn8feCFONeUUVJJ+6etH80fR83vxyDxfzyxq8DzVZtDkNq83x9Xdo1YIe0a6pdTk4IaLaOrYlW9KqBWcD48N5bFws2kvpOYQQoi4abez91VdfZf78+YwaNYr58+fz2muvATBz5kwOHjwIwEsvvcSePXuYMGECkydPJioqinvuuaexuihEs7VqRzpXxmK5hRUcuzTt2JC2J59X79050p9pVconZeaU8ud5u0nLLGzwflSVnlXE9sPn1fbdt3dEo3FdwHTP8I50jvS/6vigruE8OlaCMyFE/TXaGrQOHTrw3XffXXX8o48+Uv8/MjKSzz77rLG6JMRNIa+ogi0HK0dwosJ9OXXeMZ2w7VAmsW0DGuzeiqKw7WBlIDQwPoLBCREE+3nyyfJkrDaFwlILb3+1lycnxtEjJqTB+lLVwk0n1YLo3ToEVSuI7gp6nZanJscz5/Pd5BQ6SkANTojg4TGd0bowEBRC3Lrce/WqEOKGftx1pnIqr2ULHrizk/ra7mPZVJidK0/kjPSsYjIuOtZZGfVaenVyBGD9uoTx+3t74G1y/A1ottr59w8H+XHXjZO81teR03kcPJkDgAaYdnvDrMPz9TLy+/u6MyAujLtv7yDBmRDCpSRAE6IZKy6zVEvrMG5AFO0ifIkI8gIcSVr3pmY32P23HaocPevZKQRPj8pB+Zg2/rw8ozeh/p4AKMDX61JZ8GNKg23kURSF7zdWlnQa2DW8xoLorhIW4MXMCXGM6d9WgjMhhEtJgCZEM7Z+z1kqLhXwbhXiTULHIDQaDQPjw9VzqgZRrmS12dmRXHV6M/yqc8IDvXhpRi81ySvA2j1neX/RwQYZ2XMURHesd9PrtEwe3P4G7xBCCPckAZoQzVS52cqPuyunDMdVGcXp3yWcy+M5yadyySuqcPn9D6flUlhqAcDPx0iXtoE1ntfCy8gf7u1B706V68/2pV7kr1/vVTP8u4LVZmfhTyfV9oherQjyu35BdCGEcFcSoAnRTG36+Rwl5Y5M9iH+JvrEhqqvBfmZ6Hxpc4CiOHZaulrVkbkBceHX3bloNOiYNTme0f0i1WNpmUW8MW835y66JlfYjzvTycp1lF/y9NAzbkCUS64rhBBNQQI0IZohi9XO6p3pantMv7ZXlRMaEFdlmvPgeZeWXyott7Av9aLarml680pajYZ7hnXkwTtjuLxc62JBOW9+uYcjp/Pq1Z8Ks42vVx9V22P7RzpVEF0IIdyNBGhCNEPbDmWSX+yYHvTzNjKo69UBUq9OIRgvlVrKuFhCelaxy+6/+1i2Whg9MszHqYX4w3q25td3JeBxqSZmaYWV//v2Z7YezLzBO6/tx91n1Glcfx8jd/RuU+drCSGEO3CrWpxCiBuz2e2s3F45ejaqbyQG/dUFwD099PTsFML2w44KA9sOnadtuGvq4W07WDVzfsR1zqxZQodgXvhFT/7x/X4Kis3Y7AqfLD/Cpp/PoatDktfLed/AURDdow4F0YUQwp1IgCZEM7P7aDYXLpVP8jbpGdq95TXPHRgfrgZoO5LPc/ewDujrWbz7Qn4ZKWcLAMe0Zb8uYXW6TttwX/40ozf/+G4/Z7Md69COZxTUq2/hgV4MTnA+YBRCCHcjU5xCNCOKorA86bTaHtGrdbXcY1fq0jYQPx8jAIWlFg6nXV0z01nbq2wOiG8fiJ+3sc7XCmxh4sUHehHfvuYdoM7QajXcPzL6qrV4QgjRHMkImhDNyIETOZzNdqwl8zDobrjWSqvVMCAunFU7HFOiSYfP061j3YuGK4pSbfdmbTYH3Iinh57f3t2NjOwSisssdb5Opw7BaKwNVzVBCCEakwRoQjQTV46eDe3eslY7FQfGVwZoe1MuUlpuwctUtx2OJzIK1elVTw893esR7FWl0WhoHVq/jP8hAV5kZxfd+EQhhGgGZC5AiGYi5Uy+ukZLp9Uwqm/kDd7h0DrEh8gwR/BjtdnZfazupZ+2HarcHNCncyhGWYwvhBANQgI0IZqJqqNng7qGE+DrUev3Vt1pua2O6SwsVhs7j1yocs36T28KIYSomQRoQjQDp88XcejSAn+NxpGY1hn9uoSpZaBSzhao05TO2H88h9IKR+WCYD8T0a39nL6GEEKI2pEATYhmYHnSKfX/+3QOJSzQy6n3+3kbq+2U3F6HAupXbg7QaJzPVyaEEKJ2JEATws1l5pSwp8q6sbH9nRs9u6zqlOS2Q86VfiosNXPwZI7aHiDTm0II0aAkQBPCza3cns7lUCqhQxCRYXWrBtC9Y7CaM+1CfhknMgpr/d6dyVnY7I5edGzlR1iAcyN4QgghnCMBmhBuLKegnKTDlVOL4wbUbfQMwGjQ0adzqNquuiPzRlyd+0wIIcT1SYAmhBtbvTNdHbmKaeNPdGv/el2vanC188gFLLVI7JpxsUStdanXaegTG3qDdwghhKgvCdCEcFOFpWZ+2n9Obddn9Oyy6NZ+BPuZACitsLL/eM4N3gFJVUbPuncMxruOSW6FEELUngRoQriptbvPYLbaAYgM8yG+Xf3rVWo0mqs2C1yP3a5Um2Ktmk9NiMamKAqWkzuxnNiJYrc3dXcajb04B/OhtdgunHRqc4+oG3vBeSzHNmPLOdOk/ZBST0K4odJyC+v2ZKjtcQOiXJbWYkB8OEu3ngLg4MkcCkvMtLhGwfOj6XnkFVUA4ONpcElRcyHqyrx/Oead3wOga9UF07An0HrVb9rf3dnyzlG69A2oKAFA6x+BPnoQhugBaH2Cmrh3Nw+logTLiR1YUrdhzzruOGjwxOf+v6Hx8G6SPkmAJoQbWrHtFGWXksKGBXrRKybEZdcOC/CiYys/jmcUYLMr7DiSxchrFF2vOsLWv0sYep0MuoumYcs5g3n3osp2RjKlC1/BNOwJ9K3jm7BnDcdekkfZir+pwRmAPT8T867vMe9aiK5VLIboQejb9UJjMDVhT5snxW7FduYQlpQtWE//DHbrFSfYHJnBm4gEaEK4GbPFxpJNJ9T22P6RaLWu/ZAYGB+u1vXcduh8jQFaudlaLf/awK6ye1M0DcVmpXzjf8FefVOLUlZI2Yq/Yew+DmPvKWi0N8+vNKWihLIV76KUOCqIoDOCVguW8stnYMtIxpaRDFvmoW/XG0PMIHQtO6PRyB9S16IoCvacdCwpW7Ge2I5SVkO6IY0OfWQCxp4T0RibLqXQzfPdLMRNYsvBTPKLHdOKAb4eDIhzfWDUJzaUBWtTsNoUTp8vIiO7mFYhPtXO2ZuSTYXF8QuxZbA3beuYf02I+jLvW4r98nognQGPgb/AvHsRSpnjjwzzz8uxZh7Dc/gstL7BTdhT11CsZspW/xN73lnHAY0Oz1HPoguPxpq2B0vqNmxnD8PlDInWCqypW7GmbkXjHYgheiD6mIHo/Fs22dfgbuyl+VhTk7CkbK18rlfQhrRzPLuO/dGamv7zTgI0IdyI2WJj5fZ0tT26X2SDTCt6mwx07xjM7ksjZNsOn+fu2ztWO0dKOwl3YLtwEvO+RLXt0ecujLG3o2/bg/KNH2E7ewgAe9ZxSha+gmnoYxja9Wqq7tabYrdRvv5DbOdT1GOm2x9Tp3EN0QMxRA/EXpKHJTUJa+oW7HmVu72VklzMPydi/jkRbUh7DDEDMXToj8bkc9W9bnaK1Yz11F4sqVsd3yc1bLDQeAdg6DgAfcwgdAGtmqCX1yYBmhBuIjOnhLmLD5NT6JjC8PE0MKRbw/0FPDA+Qg3Qth/OYtqQDupUam5hOUdO5QGgwbH+TIjGpljNlG/8GBTHjk1dRCcMXUcCoPXyw3PM7zDvX4l510LHOeZSyn98D1uXEXj0n45GX/PmF3elKAoVW77Eemqvesyj/70Yogdeda7WOwCP7mMxdhuD/eJpLKlbsR7fjlJepJ5jzz5JRfZJKpK+RusfgeOnufFV6LVYrY2/69ZelF1lSrgKvRF9VK9LU8Jd0Gjdc0pYAjQh3EDSofPMW31MnVIEmDAoCg+DrsHuGd8+EB9PA8VlFvKKKjiankeXKMcuze3JWWp5qdioAAJbyAJk0fgqdv+APf/S6JDeA9PQx6qtr9JotHh0H4c+ohNl6+aiFDvy+lmS12HLSsVzxNNo/ZvP2knznsVYjm5U24aE0RgTRl/3PRqNBl1IFLqQKJR+07GdOehY9J7+c+WaPbsNe27N03qNwdxkd65OF9EZQ8wg9O16ozF6NnV3bkgCNCGaUIXZxlc/prDlYGXZJb1OyxNTutKrQ8OmtNDrtPTvEsbaPY4P7m2HztMlKhBFUaS0k2hy1vOpWA6sVtse/aejbVFzFQtdWEe8p82hfNOnWE/tAcCek07JD7MxDZ6BIWZQo/S5PszJ6zHvXaK29R0H4NHvHqeuodHp0Uf1QB/VA6W8+FLaiK3YL5x0dXebDY1fGIboQRiiBza79YkSoAnRRDKyi5m75DDnLlZuoQ8P9GLWpDh6xbckO7voOu92jYFdw9UAbc+xbB6400pWbpnaJw+Djp4uTPEhRG0olgrKN37E5UXwulZxGGKHXfc9Gg9vTCN/iSV5PRXbvwabFayO61jPJWMa9KDbpqKwnNxFxZYv1bauTVdMtz9Wr92YGpMPxrgRGONGYC/ORakodkVX6yQgwJu8vJIbn+hiGr0RTYuwZrt+VgI0IRqZoihsPpDJgh9T1EoBAAPiwnlwVAwmY+P9WLYN86VlsDfnLpZQYbGxNyVbrbsJ0KtTSKP2RwiAip3/Qym84GgYPDENfbRWv2Q1Gg3GuBHowjo6pjwLHCPB1pStlGadwHTH0+iCIhuy606znjtK+fr/cDkY1Ya0w/OOZ1yaMkTrEwg+TZdk2iPEF5224f/gvNm458o4IW5SZRVWPlqWzOcrj6rBmVGv5ZGxnXl8fGyjB0NXln7avD+THclZalumN0Vjs2YkYzm8Tm2bBt7vdMZ8XXBbvKe+ij66cmrTXnCe0sVzMB9e5zblkmw56ZSt/qeaIFXjF4bn6N+67UifaFzyp7EQjSQ9q4i5iw+RlVemHmsV7M2syfG0Cm6aUiLg2KG5cOMJFODYmXz1eICvB50jA5qsX+LWo5jLKN/0idrWRXZHHzO4TtfSGEx4DpuJpVUXyrfMA2sF2KxUbP0SS/L6ynJJ3k3zPW4vyqZsxbtgcXweaLz88Rr7e7SeLZqkP8L9SIAmRANTFIUN+zL4Zt1xrLbKKc0h3SK4746YBt2pWRuBLUx0bhvAkdN51Y73jwtzeQUDIa6nIulrdScmHt6Yhjxc7/VDhphB6ELbU7buAzXZrT0vA/PO/2He9Z1jfVvMIPRRPdHoPer7JdSKvayQ0hXvqol2MXjiOeY5tL6y3lNUkgBNiAZUWm7h85VH1XxjAB5GHQ+N6kT/BqgQUFcD48OvCtAGulH/xM3Pmr4fy7Gf1LZp8AyXFULX+kfgNelPVOxaiOXIBrBeSvygKNjOHnIkMTWYMLTvgz56ELqImAYrl6RYyilb9Q91fRxavaNKQFDN9XDFrUsCNCEaSFpmIXMXH+JiQWWixMhQH2ZNjic8sOnqu9WkV6cQvlxzDLPFMcLXNtz3qtJPQjQUpbyY8p8+U9v69n0wdOjn0nto9EZMA+7Do9dkrKf2YEnZiu3ckcoTLOVYjm3GcmwzGt9gNWO/1s91f6goditlP/4be/bltBcaTMOfRN8y1mX3EDcPCdCEqEFZhZW0zEJs9rotJk7PKmLx5rRq7x/WsxX3Du+IQd+0U5o1MRn19O4UquY/GySbA0QjKt/2FUqpY/2jxrMFHoNnNNi9NEZPDDGDMcQMxl50EctxR31GdUQLUIouYt67FPPepWjDOjryaHXoi8aj7mtFFcVO+aZP1dJUAB6DH8TQvk+9vh5x85IATYgrVFhsvPHlnmr5yerD00PHI2Ni6d255iSb7uLu2ztQVmHF18vI7T3cqyadu1KsZiwpW1Aq6v69ovUJQt++DxqdwYU9uz7FZsF6chd5KcVUlFTU6Roagyf6tt3rnfzTkrYb6/Ekte1x28ONVqha6xuMR48JGLuPx559EkvKViwndkCVf0971nEqso5TkfQV+rY9yGvTkYpS53Pj23MzsJ7YrraNPSdh7DLcJV+HuDlJgCbEFZZuTXNZcNYuwpcnJ8UT6u/+ZUX8fDz41bSEpu5Gs6EoCuXr5mI9va/e19IeWI3niKcapSyRveA8ZWvnYs85TQ1VCp1SsW1+vcrn2MsKqdj8hdrWRw/CENWznr1ynkajQRfaAV1oBzwG3Ic1fT/WlK1Y0w+Acqlcks3qCGpP7qr3/Qydb8fYa3K9ryNubhKgCVHF2exi1uw8o7Y7tvbDVIddlhqNhs6R/ozs0wa9TtIN3oysKVtcEpwB2HNOU7LoVUdZohoKY7uKJXWbI+VETQWk68iWeRRb5lHY8iX6drUvQK0oChWbv1CLe2u8AzANvN9l/aorjc6AoV1vDO16Yy8rxHpiB5aUrdgvnnLJ9fVte+Ax+MFmm91eNB4J0IS4xK4ozFt1TF03Ft3aj+d/0ROtfJCKK9iLcyjftkBt69ok1ClDvWIpw3JkkyNRqaWc8g3/xZpxuSyR61I+KJYKyrfOx5qyufKgVk+LniMpt9ZlTaSCLSfdsZ7qctJXmxnr8SSsx5PQePljiB6IPmYQuoCap8utJ7ardTMBTEMerdcar4ag9WyBMX4kxviR2HIzsJ7eh5fBTmkdpjjBkevMEDsUjdb91qEK99NoAVpaWhovvPAC+fn5+Pv78/bbbxMVFVXtnP/3//4fx44dU9vHjh3j/fffZ8SIEY3VTXEL27z/HMczHHmJdFoNM0Z1kuBMXEVRFMo3fVqZYNQvDM+Rz9Q5h5ah0xDK1n2AUuCo4GBN2ULphROYRjztktQLttwzlK+diz3/nHpM0yIMzzueIji2a71qvtpL87GmOhbZ2/POqseV0nzM+1dg3r8CbXCUYwq0Qz81Cau9JI/yKrUnDZ1vR9+ma5370Rh0ga3QBbYiMMQXWyPUyRVCozRSzYsZM2Ywbdo0Jk2axJIlS1i4cCHz5s275vlHjx7loYceYvPmzRiNxlrfJyenGHsdd97VRkiIb6MUsb6ZNIdnVlhi5uWPtlNS7ii5Mm5AW6YN7dBk/WkOz8zdNNYzMyevp2LLpc8ujQavCS+hC4+u1zUVcxnlW7/Emrqt8qBOj8eA+zHEDqvTdJiiKFiObKQiaQHYLOpxfccBmAbPQGP0dNkzUxQFe046lpStWI8nqdOW1Wh06CMT0McMwnL0J2xnDjgO+wbjPe11p9evNRX52XSePLOaabUagoKunc6oUUbQcnJySE5O5rPPHHluxo8fz+uvv05ubi6BgTUXcP3++++ZMGGCU8GZEHX17fpUNTgL9jMxfmBU03ZIuCV74QUqtn+jto0JY+odnIEj9YPnsCeqlCUyO8oSbZmHLSMZ05BHnJr+UypKKN/8OdaqC9r1RkyDHkQfM9jl6580Gg264Lbogtui9L8H25lDWFK2YD39s1pnEsWG9fS+q9btmYY+1myCMyEaU6MEaJmZmYSFhaHTOebddTodoaGhZGZm1higmc1mli1bxueff94Y3RO3uORTuSQdriwQ/uCoTk1efkm4H8Vup3zjx2oWem1AS5fvxDPEDEYb2t4xJZnr2KxiTdtNycVTeI54Cl3ojUd1bRdOOqZMiy6qx7SBrR1TpgEtXdrfmmi0evRtu6Nv2x2logTLiZ1YUrdizzp+1bmG+JGSpFWIa3DLTQJr166lZcuWxMY6/4N7veFCVwkJaZwcPTcTd31mZouNBWt3qO3B3VoyvF9U03WoCnd9Zu6sIZ9Z/o6lFJ9PcTQ0WiKm/AaPiCDX3yikE/aot8ld+wWFe1cDjsSppUvfJPD2+/HrP7HGMkSKYqdgxzKKNnwFdpt63LfnnQTd8TDaa2w6aNjvM19oPRGGTsSSe46iA5soPrQJa0E2Hq1iiBj7yDX75c7kZ9N58syc1ygBWkREBFlZWdhsNnQ6HTabjQsXLhAREVHj+QsXLmTatGl1upesQXM/7vzMFm8+qeY88/TQMfW2dm7RV3d+Zu6qIZ+ZLe8cpRu+UtvGHhMo1IdCQ/4b9b4PU2BHyn/6FMxlYLeRu/5LClL3Y7r9cXXBPTjyiZVv/Fhd1wWAwRPT0EegfV9y8s3A1TsPG/f7zBfivKwyMwAAIABJREFUxmPqMhalJB+NV4tr9sudyc+m8+SZ1exGa9BqnaDpmWeeYe3atVgslhuffIWgoCBiY2NJTEwEIDExkdjY2BqnN8+fP8+ePXuYMGGC0/cRwhmZOSWs2H5abU8d0gF/n+b317xoWIrdRvnGj8DmWEulDWqLsUfjfD4Z2vfBe9octKHt1WO2MwcoXfgK1kt1JK3njlC68JVqwZk2pD3e0+ZgaN+3UfrpDI1Gi9YnEI3WLSdwhHAbtQ7Qevfuzfvvv8/gwYOZPXs2e/fudepGr776KvPnz2fUqFHMnz+f1157DYCZM2dy8OBB9bxFixYxbNgw/Pz8nLq+EM5QFIX5a1Kw2hyjre0ifBkm5Y1EDcz7V2DPTnM0tDpMwx5Ho2u84ELrG4LXxJcwdhurHlNK8ylL/Ctlq/9J2fK/qnUsAQwJo/Ga+BLaFiGN1kchhOs5nWYjNTWVpUuXkpiYiMFgYOLEiUycOJHISOeTNDYEmeJ0P+74zLYdyuTjRMcIhEYDrzzUh7bh7rNGwh2fmbtriGdmy0mndNFr6pouY9+78Og+3qX3cIb1zAHKN3xUYxoLjckX0+0z0UfWvlyXfJ85T56Z8+SZ1cxlU5yXRUdH89xzz/HOO+9gMpl4//33mTJlCg8//DBHjx6tV2eFaAzFZRa+XV+5o2xk7zZuFZwJ96DYrI6pzUvBmTa0PcaEMU3aJ32bBLymzUF3xc5HXctYvKbNcSo4E0K4N6fG6U+ePFlt9GzSpElMmjSJwMBAFixYwNNPP8369esbqq9CuMT3G49TVOpYSxng68Hk29o1cY+EOzLvXYI951JdVp0Bz9tnukWJHq13AJ5j/4Dl4CosabvRR/XGmDD6hrUvhRDNS60DtKlTp5KRkcHYsWN599136datW7XXH3nkEb788strvFsI95ByJp+f9meq7V+MjMFklMXKojrbhZOYf16utj363oXWv+Zd501Bo9Vi7Da22ro0IcTNpda/mZ544gmGDx9+3cz+Mnom3JnVZufL1ZW1Xrt3DKZnjCykFtUpVrMjIa1iB0AX0QlD/Mgm7pUQ4lZT6zFxHx8fMjIyqh07efIkW7dudXmnhGgIq3emk3Ep55mHQccvRsY0cY/EZYq5DPPRTVjPHUG5FBg1lYrdP1QWFtd7OEoR1ZAYVgghGlKtR9DmzJnD/Pnzqx3z9vZmzpw5rF692uUdE8KVsvPLWLb1lNqeNLgdQX6mpuuQUCnlxZQuexN7niMo0vgEYYgeiCF6EFr/8Ebti/V8CpYDlZ9nHv3vRdsitFH7IIQQ4ESAlpOTQ2ho9Q+q0NBQsrOzXd4pIVzpcs4zs9UxMtMm1IeRfVo3ca8EgGKtoGz1P9XgDEApzsG8bxnmfcvQhnbAEDMIQ/u+aEwNW8ZNsVQ4pjZxpOnRtY7HEHt7g95TCCGupdbj9m3atCEpKanasR07dtC6tfyiE+5t97FsDp7MAUADzBjdCZ3seGtyit1G2dq52LJSKw96eFc7x37hBP+/vTuPjqLK+wb+reolnYTsZOkkQCCQEJbIEhAI+yKgUWAEFxRHfcVxeFFn5hHhGRXcZhwc3xmfB1FGmVHcxl2QiCzCAAqyIyEkEAiBsIQkZIGEpJda3j8aKkS27pB0dZrv5xzPyb1dXfXra5/kSy332n98H7Uf/g71a96AdHQ3VEVqkXrsWz+DerbM1TAHwjL0IQiC0CLHIiK6FrfPoM2cOROPP/44Jk+ejHbt2uHYsWP46quv8Oc//7kl6yO6LnU2CR9/X6C1h/dOQHI8V6nQm6qqsP+wBHLxz1pfwMB7Yeo2ElJxDqSCHyEV5wDq+UW/FQlS0Q5IRTsgWEJg7DwAppRMiFEdmiVESSfy4Mxbq7Utg+6D2KYFFkInInKT2wFt9OjR+Ne//oUvvvgCGzZsQFxcHBYvXoz0dE6MSL7r642HcabWtRhzWLAZdw7rdI13kDc4dnwF54GNWtt8060w9xwLADB17AtTx75QbDWQDm2B8+DmhqWWAKi2Gjhz18CZuwZiRAJMKZkwdh4IRDdtsmHVUQ/bhn9qbWOH3jB2yWziJyMiah4eTQCVnp7OQEatRlHJWazbdVxr3zOqC4IsJh0r8l2qox5S0Q44i3ZADAyFuf8UiIGhLXIsR+73cOxerrWNXTJh7j/lku1ESwjMPcbA3GMM5KoTkAo2wXlwc6N1J5WqE7Bv/Qz2bZ/jeHQHSE14AFS1n4Na67oEjoBgBAz5NS9tEpHuPApo+fn52LFjB6qqqnDxEp5PPvlksxdGdD1kRcGSlftx4VvavWMk+qfxabyLqYoC+WQenAWbIBXtBGTXmUYZgHRsLywjHoUxoVuzHtN5eBvsmz/S2oZ26bAMu/a9XoaIBBhuvgvmfpMvWzNUFY6yI9ddn2XwryEGhV/3foiIrpfbAe3TTz/FK6+8gszMTGzcuBFDhw7Fpk2bMGrUqJasj8hjiqri/ZUHUFxaCwAwGUVMuyWFZ0XOkytPQDp46dmoi6l11aj/9q8w97kd5j4TmmWJI+lkPmzr3saFpyTFmE4IHP1/IYju/ztREEUYE3vAmNij4axfwSbIJde/DrAxZQhMyf2vez9ERM3B7d+MixcvxuLFi5GRkYF+/fph4cKF2LBhA1asWNGS9RF5RFFUvPtdPjbtPaX13T4oCTERQTpWpT+l/iykwq1wFmyCcvrIZbcRIxNh7NAbzvz1UG01AFQ4dn0D+eR+WEY+BrFNZJOPL58+ivpV/wOcfwJTDItD4LjfQzAFNHmfgjkQptQhMKUOgVJXjXCTHVXVdU3bmTHA63OuERFdjUfzoGVkZAAARFGEoigYNmwYZs2a1WLFEXlCUVS8uyIfm3IbwllmzzjcOqCDjlXpR5WdkI7+DOng5sZPRF5ECAyFsfNAmLoMgqGta5xM3UbC9p+3IZ/MBwDIpwpQ9+VcWIY/AmOHXh7XoZwtR/13fwOcNtcxg8IReOtTEC1Nu6n/csSgcAREh8Bgqmm2fRIR6cntgBYXF4fjx48jMTERSUlJWLt2LSIiImAy8aZr0p+iqPjnt/n4aV9DOBucbsWD47tCbGWXNlVFgbO6DEpNbdPef64azkM/wVm4FbCfu3QDgxHGDn1gShkEQ2KPSy4xisERCLx1Fhw/L4dj51JAVaHaa1G/6nWYetyCgJvvgmBw71eHUn8WdSteg1p/xtVhDkTgrf8FMaRtkz4bEdGNwu2A9sgjj6CwsBCJiYmYMWMGnnzySTidTjzzzDMtWR/RNcmKgn9m52NLXqnWN/QmKx4Y1/rCmVJXjbplL6O25nSz71uM7QxTymCYOvWD8IsJYX9JEEUE9JkAg7UrbOsWQT1XBQBw5q6GfKoAgaNnXHMJJNVpQ/3Kv0M9e/7/i8GIwLG/gyGyXbN8HiIif+ZWQFNVFf369YPVagUADBs2DNu2bYPT6URw8NV/0RO1JFlRsDg7H1svCmfDesVj2tjUVhfOAMCx5zuozRjOhJC2DetahsV6/H6jNRXBd76E+vXvQC7eAwBQTh/BuS/nwjLkQZg6D7js+1RZQv2aNxrmLxMEWEY+BqM1tcmfhYjoRuJWQBMEAbfffjt27dql9ZnNZpjN5hYrjOhaZEXBO8vzsC2/TOsb3jsB99+S0irDmeqog3P/Bq0tBEcAQhOWpBKNMFpTYEwZDENcFwhN2cdFBEsbBI79HZy5q2Hf+hmgyIDTBtu6RZBP5iFg0H0QjA03+6uqAtuGf0I+nqv1BWQ+AFPHjOuqg4joRuL2Jc60tDQUFRUhOTm5JeshcoskK3h7eR527G8IZyP6JOD+Ma13Og1n/gbtRnpT20QETHrxusNVcxEEAeaeY2GIS0H92re0NSud+zdCLi2EZdQMGCITAAD2LZ9COtSwbq+5zwSYu43QpW4iotbK7YDWv39/TJ8+HZMmTUJcXFyjP4KTJ09ukeKILkeSFfzjm33YeaBc6xvVJxFTx3RpteFMVSQ4ctdo7bD+t8PuI+HsYobojgj+1Quw/fAepMKtAFyz+dd9/QICBk0FHPVw7l2lbW9KGw5z34l6lUtE1Gq5HdB27dqFhIQEbNu2rVG/IAgMaOQ1kqxg0bJ92FXQEM5GZyTi3lGtN5wBgFS4Deq5SgCuqS/a9BwKe5Vd56ouTzAHwjLyMTgTusG+6SPXbP6yA/Yf3mu0nTGpLwIyH2jV/1+IiPTidkD74IMPWrIOomuSZAVvLc3F7oMNN9GPyWiHe0Z1btUhQFVVOHJWam1T99EQjWYAvhnQgPOXPLsOgyGmM2xrF0KpOtnodUNcCiwjfwNB9L2zgERErYHbvz0VRbnif0QtzSkpePPrxuFsbP/WH84AQD6ZD6Wi2NUwmGHuNlLfgjxgiExA0KR5MHUdqvWJkYkIHPskBCMfIiIiaiq3z6B169btin8I8/Pzm60gol9ySgoWfr0XOYUVWt+4m9tjyvDkVh/OADQ+e5Y6GIKljY7VeE4wBsAy9GEYOw+EUnkcppRMCOYbe2ktIqLr5XZAW7t2baN2eXk53n77bYwYwaezqOU4JRkLv85tFM7GD2iPycP8I5zJlScgH8s533I9KdlaGePTgPg0vcsgIvILbge0hISES9rz58/H5MmTMWXKlGYvjMgpyVjw1V7kHq7U+m4b2AG/GtrJL8IZADj3Npw9Myb1adJkskRE5H+u6w7e2tpaVFZWXntDIg+pqop3svMbhbOsQUleC2fOQz+hftX/QDq5v8WOodRVw3nwovnC0se12LGIiKh1cfsM2qxZsxr9YbTZbNi+fTvuuOOOFimMbmzrfz7ZaBLaOzKTMGFwR6+EM6X6FGzr3gagQjqRh+DJL11z3cmmcOZ+DygSANc6mYa4Ls1+DCIiap3cDmgdOnRo1A4MDMQ999yDQYMGNXtRdGM7VlaLf39/UGsP752AiUM6ee34jr2rAKiuhmSHbcM/EZg1u1ln9Veddjjy/6O1W/O9Z0RE1PzcDmgzZ85syTqIAAB2h4xFy3Ihya7pWxKj2+DeUZ29dnyl/iycBT826pNLDsCZ+z3MPW9ptuM4D/wA2M8BAISQaBiT+jbbvomIqPVz+5TAyy+/3GixdMC1usCf/vSnZi+Kblwff1+Akoo6AIDZJOKxCd1hMhq8dnxn3n8A2elqiA3/frFv+xxKdUmzHENVlPNn6VzM6WM5oSsRETXi9l+F7Oxs9OjRo1Ffjx49kJ2d3exF0Y1pa14pfshpCEH3jUlBfNtgrx1flRxw7vtea1uGPggxsp2rITtRv34xVEW+7uNIR3ZCrTm/VFVAMEwpQ657n0RE5F/cDmiCIEBV1UZ9sixzJQFqFmXV9ViysuGJyQHdYjG4p9WrNTgPboZqqwEACG2iYOw8AJYR0wHRdQZPKSuEI+e76zqGa1mnhn2Yu42EYAq4rn0SEZH/cTugZWRk4PXXX9cCmaIoWLBgATIyMlqsOLoxSLKCfyzLhc3hOjsVEx6IaWNTvTrXmaoqcF40o7+5xxgIohGGqPYw95mg9Tt2LIVceazJx5FLD0EpO+xqiEaYuo9q8r6IiMh/uf2QwDPPPIPf/OY3GDx4MOLj41FSUoLo6GgsWrSoJeujG8BXGw6jqMR15sogCvjNhO4IDHD7q9ks5OI9UM6ccjVMgTB1Haa9Zu51G6Sju6GUFwGKBNt/FiNo0nMQRM9rvDgEmroMhBgUft21ExGR/3H7L0xcXBy+/vpr5OTkoKSkBFarFenp6RB5czNdh5zCCqzcVqy1Jw9PRkdrqNfraLQeZtowCOZArS2IBliGT0fdV3MBWYJScRSOXcsRkDHJo2MoZ05BOtLwoI2pJyemJSKiy3M7XeXn56O0tBS9evXC+PHj0atXL5SWlmL//pabaZ38W1WNHYuz87R2enIUxvRr5/U65LLDkEsOuBqCAeYeYy7ZxhARj4B+d2ptx+7lkMuPeHQcx97VuDC/mqFdOgyRCVd/AxER3bDcDmizZs2CJEmN+pxOJ2bNmtXsRZH/UxQVi7PzUFvvmtIirI0ZD9+WBlGHNTYvPntmTO4PsU3UZbcz9RgLQ1yKq6EqsK1/B6rkcOsYiq0GzgMN86txWSciIroatwPayZMn0a5d47Mb7du3x4kTJ5q9KPJ/3245ivyjVQAAAcCjt3dHaJDZ63UoNeWQirZr7asFJ0EUYRn+CGB01alUnYBj51K3juPMWwfIrjAnRnWAIT7tOqomIiJ/53ZAi4uLw759+xr17du3DzExzb9GIfm3g8erseyHIq2dNSgJaR0idKnFsXcNcH76GENCNxjadrjq9mJoDAJuvrvh/TnfQT518CrvuDC/2lqtbU4f69UnVImIqPVxO6A9+OCDmDFjBj744ANs2LABH3zwAWbOnImHHnqoJesjP1Nb78Tb3+yDcj4UdUkMwx2Dk3SpRbWfg3P/Bq3t7mVHU7cRMCR0O78TFfUbFkOV7Ffc3nnoJ6j1ZwEAQnAkjMn9m140ERHdENx+ivOuu+5CSEgIvvjiC5w6dQpWqxWzZ8/GuHHu/VErKirCnDlzUF1djfDwcMyfPx9JSUmXbLdixQq89dZbUFUVgiDg3XffRdu2bd3+QOS7VFXFe9/tR8VZV5gJthjxmzu6w6DTk8CO/PXA+WAlRiTAkNjTrfcJggjLsP+Dc58/AzhtUM+Uwr7tC1gG3XfJtpfMr9ZzTJOm5yAiohuLR38p+vXrB7PZjKoq171DtbW1+OKLLzB58uRrvnfevHmYOnUqJkyYgGXLlmHu3Ll4//33G22zd+9evPHGG1iyZAmio6NRU1MDs9n79yVRy/jP7hPYVVCutR+6NQ2RoRZdalFlCc7cNVrbnD7Oo8uOYpsoWAZOhW3jvwAAztw1MCb1gfEX95bJx3Ia1vA0WRrNr0ZERHQlbp+6+P777zFmzBgsWLAAzz//PD788EPMmzcPy5Ytu+Z7KyoqkJeXh6ysLABAVlYW8vLyUFlZ2Wi79957Dw8//DCio6MBACEhIQgI4DI4/qC4tAafrD2ktUf1SUSflGjd6pEKt0KtqwYACIFhMHYe4PE+jKlDYGh/k9a2rV8M1VHfaBvHnovnVxsOwRzUxIqJiOhG4nZAe/311/HnP/8ZS5cuRWBgIJYuXYoXX3zxkgXUL6ekpASxsbEwGFxrGhoMBsTExKCkpKTRdoWFhTh27Bjuu+8+TJo0CW+++eYl639S62OzS1i0bB8k2bVMWLuYNrhrZLJu9fxyPUxTj9EQDCaP9yMIAixDHwICXAu6q7UVsG/5RHtdLj8CueT8PIFXmF+NiIjocty+xHny5EmMHz++Ud+kSZOQmZmJ2bNnN0sxsizjwIEDePfdd+FwOPDII48gPj4eEydOdHsfUVFtmqWWq4mODmnxY/iT//lkN05V1gEALGYD/vhQf8TH6DeGdYf3oLbyOABAMAXAOuR2GAKbWk8IasdPR9nS1wEAzv0bENVrCIKSe6N00/faVm26DUJMxySP9szvmec4Zp7jmHmOY+Y5jpnn3A5oUVFROH36NNq2bYuEhATs3r0bERER2uLpV2O1WlFaWgpZlmEwGCDLMsrKymC1WhttFx8fj3HjxsFsNsNsNmPUqFHIycnxKKBVVNRCUVrurFt0dAjKy2tabP/+Zsu+U/h+e8NSTveNSUGAAF3HsO6Hr7SfjSlDUFkLoLbp9ajRN8HYMQNS0Q4AQOk3byBw/B9Ql7dZ20ZJHeXRZ+b3zHMcM89xzDzHMfMcx+zyRFG46kklty9xTpkyBTt37gTgmnLjgQcewIQJE3Dvvfde871RUVFIS0tDdnY2ACA7OxtpaWmIjIxstF1WVhZ+/PFHqKoKp9OJLVu2oGvXru6WSD7m4PFqLFl1QGsP7B6LQT3idKwIkCuOQT6e62oIAsw9b7nufQqCgIDBD0CwuP6FqNZVo275K4Dq+seLIT4NhrZJ130cIiK6cbh9Bu3RRx/Vfp44cSL69++P+vp6JCe7dy/R888/jzlz5uDNN99EaGgo5s+fDwCYPn06nnjiCfTs2RO33XYbcnNzceutt0IURQwePNitJ0TJ9xQcq8bfP98Du0MGAMREBOL+W1J1n6DVsfeiZZ2S+kIMbZ6JlsXAUAQMeRC2NQvOH6jhYQEu60RERJ4SVD+7C5+XOPV3oLgKr3+eA7vTFc7C2wTgv+7phYS2wbrWpZyrwrl/PwUorrqCJj4HQ0zzPqxQv+4fkA79pLXFiHgETX4ZguDZXG/8nnmOY+Y5jpnnOGae45hdXrNd4iRyx4HiKteZs/PhLCzYjD/PyNQ9nAGuucouhDNDbJdmD2cAYMm8H0JQuNY29xzncTgjIiLiXw5qNvlHXeHM4XTdexXWxoynp/ZGu1j9n95RHfWulQPOM93UMpcdhYBgBI59EmJsZ5jShsOYktkixyEiIv/GNWeoWeQdqcT/fpEDh+QKZ+FtzHh6ah/ERfrGxKzOAz8ADtdUH0JYLIzte7fYsQzRHRE84dkW2z8REfk/BjS6bvvOhzPn+XAWERKAp+/tjVgfCWeqIsOxd5XWNvccC0Gn9T+JiIjcwYBG1yW3qAILvtzbOJxN7Y3YCN8IZwAgFe2EWlsBABAC2sDEy45EROTjGNCoyfYedoWzC0s4RYa6zpzF+FA4u2RZp+4jIRi5visREfk2BjRqkpzC03jjq72QZNeUJlGhAZg1tQ9iwgN1rqwx+VQBlPIiV8NghKnbKH0LIiIicgMDGnns50On8ebXF4czC2ZP7Y22PhbOAMCZ0zAxralLJsSgMB2rISIicg8DGnnk54OnsfDrvZDPTwbcNsyCp6f2Rtsw3wtnSnUJpKO7tbYpfayO1RAREbmPAY3ctrugHG8uzW0UzmZP7YOoMIvOlV2eI6fhyU1D+5tgCI/XsRoiIiL3MaDRVSlny2Df8TUqahxYVxgBVYkDICImPBBPT+2NyFDfC2eqIkE+thfOg5u0PnP6eB0rIiIi8gwDGl2Rs3ArbBvfBZw2hAF4tA1wRglEntAFGeMn+FQ4U1UVSkUxnAWbIBVugVp/VntNbJsEgzVVx+qIiIg8w4BGl1AlO+ybP4Zz/4ZLXgsT6zEQOcDKHJyLag9TSiaMyQN0u/leqauGdPAnOA9uglJ5/NINBAEBGZMgCIL3iyMiImoiBjRqRK48AdvaN6FUndD6yuUQ5DkT0NdyFG1Qr/UrFcWw/1QM+5ZPYWjX0xXW2veCYDS3aI2q5IB0ZBecBzdBPp4LqOol2whB4TB1GQRj6mDee0ZERK0OAxoBcF0idB7YCPumjwDZofXvtCfh03MDYLQEIfPuJxB45hCcBT9COroLkKXzb1YgF++BXLwHMAfBlNzfNaVFbOdmO3OlqirkUwWQCjbBeXg74Ky/dCODGcaOfWFKyYQhvhuXcyIiolaLAY2gOuph++E9SIVbG/oMJnxWk4HNts4ABEwd0RkRYUFAWDqM7dOh2s/BeXg7pIObIZ8qaNiZow7O/PVw5q+HEBoLU8ogOPoOh1J36Vkut2qzn4NUuA3Og5uh1pRfdhuDNRWmlMEwdsyAYPa96T6IiIg8xYB2g5PLi1C/9i2oZ8u0PiE8Hh/VD8c2m+vrkZIYhsHp1kbvEwKCYU4bDnPacChny+As2HRJiFLPlsKx42sc3/F1s9ftCn+ZMHUZCDEkutn3T0REpCcGtBuUqqpw5q6GfetngCJr/aauQ7EteAS2rXEtj2QQBUwb1/WqlyrF0BgEZEyCue8EyKcOQjq4Cc7CK1yGvB4XLp+mDIYYk8wb/4mIyG8xoN2AVFst6te/47pn7AKTBZYhD6Le2gefv7NF6x53c3sktA12a7+CIMJoTYXRmoqAQfefv5F/M1B1DIosX3sHl92pCLFtB689gEBEROQLGNBuMFLJAdjWLYJ6rkrrE9smIXDUbyGGxeK95ftwzua6+T863ILbByU16TiC0QxT5wEwdR6A6OgQlJfXNEf5RERENwQGtBuEqihw/Lwcjp1LG01LYeo5FgH9p0AwGJF3pBI/7SvVXpt2SyrMJoMe5RIREd3QGNBuAEpdNWzr/gH5ZH5DZ0AwAodPh7FDLwCAU5LxwaoD2sv902LQo1OUt0slIiIiMKD5PVVVYVuzEHLpQa3PEJcCy8jHILaJ1Pq+/ekoSqtcN/UHBhhwz6guXq+ViIiIXBjQ/JxSdfyicCbA3Od2mPtMgCA2XLosqTiHFVuOau07hyUjvE2AlyslIiKiCxjQ/JxUuE372dipHwIyftXodVVV8eHqAkiy6760jtZQDO+V4NUaiYiIqDGuhePHVFWF8+KAlnzzJdv8tO8U8o+6nugUBQG/HpcKUeT8YkRERHpiQPNjSkUx1LPnn8o0WWBs17PR67X1Tny67pDWHp2RiPaxId4skYiIiC6DAc2PXby2prFD70smef1i/SHU1DkBAJGhAZg4pKNX6yMiIqLLY0DzU6qqwnm44fKmKbl/o9cLjlVj454SrX3f6BRYzLwlkYiIyBcwoPkppbwIas1pV8McCENiD+01SVYazXnWu0tb9E7hguNERES+ggHNT1189syY1AeCwaS1V20rxonT5wAAASYD7huT4vX6iIiI6MoY0PyQqiqNptcwdWp4erO8uh7LNx3R2hOHdERkqMWb5REREdE1MKD5IaW0EOq5SlcjIBiGxG4AGuY8c0gKAKB9TBuMzkjUq0wiIiK6AgY0P9To4YCOfSGIrpv/dxwox97DFQAAAcAD47rCIPIrQERE5Gv419nPqKoC6fB2rW3s5Hp6s84m4ePvC7T+4X0S0Ck+1Ov1ERER0bUxoPkZ+dRBqHXVAADBEgJDfBoA4OuNh3Gm1gEACAs2486hybrVSERERFfHgOZnGk1O2zEDgmhAUclZrNt1XOu/d3QXBFk45xkREZGvYkDzI6qiQCraobWNyf2hqCreX3UA6vm+Hh0j0a9rjD4FEhERkVvQMKh6AAAVnklEQVQY0PyIXLIfav1ZAIAQGAZDXCq255fh6KkaAIDJKOL+sakQBC6GTkRE5MsY0PzIxXOfGTtlQFaBrzYWan239GuHmPBAPUojIiIiDzCg+QlVkX5xefNmbPj5JMqrbQCAYIsR429ur1d5RERE5AEGND8hn9wP1V4LABCCI+AIT8I3m4q0128bmIQgi+lKbyciIiIf4rVH+YqKijBnzhxUV1cjPDwc8+fPR1JSUqNtFixYgI8//hgxMa6b2Pv06YN58+Z5q8RWrdHTm536Y+X246ipcwIAIkMDMKpvgl6lERERkYe8FtDmzZuHqVOnYsKECVi2bBnmzp2L999//5LtJk6ciNmzZ3urLL+gyhKcRTu1tj2+N1Z9eUxrTxrSCSajQY/SiIiIqAm8comzoqICeXl5yMrKAgBkZWUhLy8PlZWV3ji835NP5AKOOgCA0CYK2fsF2B0yACAhOhgDu8fpWR4RERF5yCsBraSkBLGxsTAYXGdxDAYDYmJiUFJScsm23377LW6//XY8/PDD2L17tzfKa/WchQ1LOzkSemP9npNa+85hyRBFTqtBRETUmvjUdPL33HMPHnvsMZhMJmzatAkzZszAihUrEBER4fY+oqLatGCFLtHRIS1+DHcpkgNHixuC7MYziZAVBQDQvVMURg9I8ol5z3xpzFoLjpnnOGae45h5jmPmOY6Z57wS0KxWK0pLSyHLMgwGA2RZRllZGaxWa6PtoqOjtZ8zMzNhtVpx8OBB9O/f3+1jVVTUQlHUa2/YRNHRISgvr2mx/XvKeWQXVLvr8qYc1Bbf5MkAXIFsQmYSTp+u1bE6F18bs9aAY+Y5jpnnOGae45h5jmN2eaIoXPWkklcucUZFRSEtLQ3Z2dkAgOzsbKSlpSEyMrLRdqWlpdrP+fn5OHHiBDp27OiNElutiyenzZGScCGc9UmJRueEMH2KIiIiouvitUuczz//PObMmYM333wToaGhmD9/PgBg+vTpeOKJJ9CzZ0/87W9/w759+yCKIkwmE1599dVGZ9WoMVVyQDracHlzdbnrYQBBAH41tJNeZREREdF18lpAS05Oxueff35J/zvvvKP9fCG0kXuk4j2AZAcAVCIcJ2XXvXpD0q2IbxusZ2lERER0HbiSQCsmHW64vLmtvj0AASajiAmDefaMiIioNWNAa6VUpw3S0T1ae7cjCQAwJqMdIkICdKqKiIiImgMDWislFe8BZAcAoEQKxyk5HMEWI24dwAXRiYiIWjsGtFbq4qc3d50/e8YF0YmIiPwDA1orpDrqIR1ruLz5s6MDF0QnIiLyIwxorZB0dDcgSwCA41IEypQwTBzMBdGJiIj8BQNaK+S86PLmbkcSEtoGY1APLohORETkLxjQWhnVfg7Ssb1ae7cjiQuiExER+RkGtFZGOrILgioDAIqlKETGJ+KmzlE6V0VERETNiQGtlTmb/5P2825HB0wZ3hmCwLNnRERE/oQBrRVRbbUwlu3X2o74PuicyAXRiYiI/A0DWitydMdGiFAAAEekthg3srfOFREREVFLYEBrJRRVRW3+Zq1dHZXOBdGJiIj8FANaK7E75zDaKSe0dvdhY3SshoiIiFoSA1orIMkKCreshyioAIDKwHaIjLPqXBURERG1FAa0VmDN9mPoLBVo7aieg3WshoiIiFoaA5qPO11dj42bcpBsLAUAqBAQlHKzzlURERFRS2JA82GqquKzVTl4KPB7XFgowGBNhRgUrm9hRERE1KIY0HzYrryTyDz9BazGMwAAVTAgIGOSzlURERFRS2NA81F19XYoP/wDnUzlWl/giOkwWlN1rIqIiIi8gQHNB6mqisNfv4WuYrHWJ2TcBVPnATpWRURERN7CgOaDyjZ8gg61P2vt0wnD0KbPrTpWRERERN7EgOZj7LlrEVSwSmsXmNLQYfyvdayIiIiIvI0BzYc4D2+HffOHWjvfmYDEO2ZAFPm/iYiI6EbCv/w+QjqZj/p1/4AA12oBR6S2KEt/ALFRITpXRkRERN7GgOYD5Ipi1K/6XwiKBAAolUOx1Hgbxg7sonNlREREpAcGNJ0pZ8tRv+L/Ac56AMAZJRBv1YzGlLG9YDLyfw8REdGNiAlAR0r9WdR99xrUetdEtHWKCYtqRqNbt87o2iFC5+qIiIhILwxoOlGdNtSv/DvUM641Np2qiMW1I3DGFI27RnbWuToiIiLSEwOaDlRZQv2aN6CUFwEAFFXA+7VDUCjF4a4RnRESZNa5QiIiItITA5qXqaoC24Z/Qj6eq/V9UdcfOc4OSEkMw+B0q47VERERkS9gQPMy+9bPIB36SWt/V5+OTfZUGEQB08Z1hSAIOlZHREREvoABzYscOd/BmbNSa29zpmJl/U0AgHE3t0dC22C9SiMiIiIfwoDmJc6inbBv+VRrHw/ojI9r+gEQEB1uwe2DknSrjYiIiHwLA5oXqKoC+7bPtLYtvBNeL+kP9fzwT7slFWaTQa/yiIiIyMcY9S7gRiAd/VmbTgPmQLxZNQTO80s69U+LQY9OUTpWR0RERL6GZ9C84OL7zo6E9MbRKlc4Cwww4J5RXM6JiIiIGmNAa2FyWSHkUwUAAFUwYElRvPbancOSEd4mQK/SiIiIyEcxoLUwx0VnzwoMKaiUggAAHa2hGN4rQa+yiIiIyIcxoLUg5WwZpKIdWntphWsJJ1EQ8OtxqRBFznlGREREl2JAa0GOvasB1XW/WYEUj5OyawH00RmJaB8bomdpRERE5MMY0FqIaquF88APWvv7um4AgJiIQEwc0lGvsoiIiKgV8FpAKyoqwt13342xY8fi7rvvxpEjR6647eHDh3HTTTdh/vz53iqv2Tny1wOSHQBwUgrHAckKgyjgsQndYTFzdhMiIiK6Mq8FtHnz5mHq1KlYtWoVpk6dirlz5152O1mWMW/ePIwePdpbpTU7VXbCmbtGa6+zdQcgYMqIzkiKC9WvMCIiImoVvBLQKioqkJeXh6ysLABAVlYW8vLyUFlZecm2b7/9NoYPH46kpCRvlNYipENboNafAQBUK4HY5UhCenIUxmQk6lwZERERtQZeCWglJSWIjY2FweBazshgMCAmJgYlJSWNttu/fz9+/PFHPPjgg94oq0Woqgr7RVNrbLSlIaRNIB6+LQ2CwKc2iYiI6Np85mYop9OJ5557Dq+88ooW5JoiKqpNM1Z1edHRV34Cs65wN2qrTgAAbKoRPzlS8OxD/ZDc4cZezulqY0aXxzHzHMfMcxwzz3HMPMcx85xXAprVakVpaSlkWYbBYIAsyygrK4PVatW2KS8vR3FxMR599FEAwNmzZ6GqKmpra/HSSy+5fayKilooitrsn+GC6OgQlJfXXPH102s+w4W1AbbYu2D0wBTEhQVc9T3+7lpjRpfimHmOY+Y5jpnnOGae45hdnigKVz2p5JWAFhUVhbS0NGRnZ2PChAnIzs5GWloaIiMjtW3i4+OxdetWrb1gwQLU1dVh9uzZ3iixWdScKERAhWtZJ0UVUBzeH49lJulbFBEREbU6XnuK8/nnn8eHH36IsWPH4sMPP8QLL7wAAJg+fTr27t3rrTJajKqqKFzzhdbOlZNw78SBMIicao6IiIg847V70JKTk/H5559f0v/OO+9cdvvHH3+8pUtqVj9syUO6fT9w/jmAiP63IzLUom9RRERE1Crx9E4zKC6tQfWOFTAIrnvfTlvaoVtGH52rIiIiotaKAe062RwS/rVsNwaYCrQ+6+CJOlZERERErR0D2nX6aE0BOtXlIFB0AgCUNrEI6Nhb56qIiIioNWNAuw4/5Z7CT3tPYlhAvtYX2Hs8BIHDSkRERE3HJNFEpZV1eH/1AfQyH0Wk4RwAQLCEwNRlkM6VERERUWvHgNYETknBomX7YHdIGGnZp/Wbuo+CYDTrWBkRERH5Awa0JvhyQyGOltags7EU7YznF3w3mGDqNlLfwoiIiMgvMKB5aFveKazefgwAGp89SxkMMTBUr7KIiIjIjzCgeaCqxo7X/70bABArVqO7+cT5VwSYe47VrzAiIiLyKwxoHliycj9q6hwAgHGhB7R+Y4deEMPj9CqLiIiI/AwDmpsUVcWB4moAQKhYj96GQ9prppvG61UWERER+SEGNDeJgoDJw5PRKSEMT/SogKDKrv6YTjDEdtG5OiIiIvInDGgeGNU3EX+fOQDR5Vu1PnP6OAiCoGNVRERE5G8Y0DxUk7MesJ+fmDYkGsakvvoWRERERH6HAc0DqqLgzLblWtvc8xYIokHHioiIiMgfMaB5QDq6C1LVKVfDHART6hB9CyIiIiK/xIDmAUfOSu1nc7eREEwWHashIiIif8WA5gGlwrWCAEQDTD1G61sMERER+S0GNA+Ye90GY3gsLIN/DTEoXO9yiIiIyE8Z9S6gNQnocweix96H8vIavUshIiIiP8YzaEREREQ+hgGNiIiIyMcwoBERERH5GAY0IiIiIh/DgEZERETkYxjQiIiIiHwMAxoRERGRj2FAIyIiIvIxDGhEREREPoYBjYiIiMjHMKARERER+Ri/W4tTFAW/OIa/4Zh5jmPmOY6Z5zhmnuOYeY5jdqlrjYmgqqrqpVqIiIiIyA28xElERETkYxjQiIiIiHwMAxoRERGRj2FAIyIiIvIxDGhEREREPoYBjYiIiMjHMKARERER+RgGNCIiIiIfw4BGRERE5GP8bqmnllRUVIQ5c+aguroa4eHhmD9/PpKSkvQuy6eNHDkSZrMZAQEBAICnnnoKQ4YM0bkq3zF//nysWrUKJ06cwPLly5GSkgKA37WrudKY8bt2ZVVVVXj66adRXFwMs9mMDh064MUXX0RkZCR+/vlnzJ07F3a7HQkJCfjrX/+KqKgovUvW3dXGLDU1FSkpKRBF1zmOV199FampqTpXrL8ZM2bg+PHjEEURQUFBeO6555CWlsbfZ02lktumTZumLl26VFVVVV26dKk6bdo0nSvyfSNGjFAPHDigdxk+a/v27erJkycvGSd+167sSmPG79qVVVVVqVu2bNHaf/nLX9T//u//VmVZVkePHq1u375dVVVVXbhwoTpnzhy9yvQpVxozVVXVlJQUtba2Vq/SfNbZs2e1n9esWaNOnDhRVVX+PmsqXuJ0U0VFBfLy8pCVlQUAyMrKQl5eHiorK3WujFqzjIwMWK3WRn38rl3d5caMri48PBw333yz1u7VqxdOnjyJ3NxcBAQEICMjAwBwzz33YOXKlXqV6VOuNGZ0ZSEhIdrPtbW1EASBv8+uAy9xuqmkpASxsbEwGAwAAIPBgJiYGJSUlCAyMlLn6nzbU089BVVV0bdvX/zhD39AaGio3iX5NH7Xmo7ftWtTFAX//ve/MXLkSJSUlCA+Pl57LTIyEoqiaJeiyOXiMbtg2rRpkGUZQ4cOxeOPPw6z2axjhb7jmWeewaZNm6CqKhYvXszfZ9eBZ9CoRX300Uf45ptv8OWXX0JVVbz44ot6l0R+it8197z00ksICgrC/fffr3cprcYvx2z9+vX46quv8NFHH+HQoUNYuHChzhX6jj/96U9Yv349fv/73+PVV1/Vu5xWjQHNTVarFaWlpZBlGQAgyzLKysp4qeUaLoyP2WzG1KlTsWvXLp0r8n38rjUNv2vXNn/+fBw9ehSvv/46RFGE1WptdNmusrISoijy7NlFfjlmQMN3rU2bNpgyZQq/a5cxceJEbN26FXFxcfx91kQMaG6KiopCWloasrOzAQDZ2dlIS0vjKdqrqKurQ01NDQBAVVWsWLECaWlpOlfl+/hd8xy/a9f2t7/9Dbm5uVi4cKF2Oa5Hjx6w2WzYsWMHAOCTTz7BuHHj9CzTp1xuzM6cOQObzQYAkCQJq1at4ncNwLlz51BSUqK1161bh7CwMP4+uw6Cqqqq3kW0FoWFhZgzZw7Onj2L0NBQzJ8/H506ddK7LJ917NgxPP7445BlGYqiIDk5Gc8++yxiYmL0Ls1nvPzyy1i9ejVOnz6NiIgIhIeH49tvv+V37SouN2aLFi3id+0qDh48iKysLCQlJcFisQAAEhMTsXDhQuzatQvz5s1rNM1G27Ztda5Yf1cas0ceeQRz586FIAiQJAm9e/fGH//4RwQHB+tcsb5Onz6NGTNmoL6+HqIoIiwsDLNnz0b37t35+6yJGNCIiIiIfAwvcRIRERH5GAY0IiIiIh/DgEZERETkYxjQiIiIiHwMAxoRERGRj2FAIyJqRsePH0dqaiokSdK7FCJqxRjQiIiIiHwMAxoRERGRj2FAIyK/V1paiscffxwDBgzAyJEj8f777wMAFixYgCeeeAK/+93v0Lt3b0yaNAn79+/X3ldYWIhp06YhIyMDt912G9auXau9ZrPZ8Je//AUjRoxA3759ce+992pLAAHA8uXLMXz4cNx888146623vPdhicgvMKARkV9TFAW//e1vkZqaio0bN2LJkiVYsmQJfvjhBwDA2rVrMW7cOGzbtg1ZWVmYMWMGnE4nnE4nHnvsMWRmZmLz5s149tln8dRTT+Hw4cMAXIto79u3D5988gm2bduGWbNmaYtpA8DOnTuxcuVKLFmyBAsXLkRhYaEun5+IWicGNCLya3v37kVlZSVmzpwJs9mMdu3a4a677sKKFSsAAN27d8e4ceNgMpnw0EMPweFwYM+ePdizZw/q6urw6KOPwmw2Y+DAgRgxYgS+/fZbKIqCL7/8Es888wxiY2NhMBjQp08fbUFtAJg5cyYsFgu6du2Krl27NjozR0R0LUa9CyAiakknTpxAWVkZMjIytD5ZlpGRkYH4+HjExcVp/aIoIjY2FmVlZQCAuLi4RmfF4uPjUVpaiqqqKtjtdrRr1+6Kx714wfHAwEDU1dU158ciIj/HgEZEfs1qtSIxMRGrV6++5LUFCxbg1KlTWltRFJSWliImJgYAcOrUKSiKooW0kpISJCUlISIiAgEBATh27Bi6du3qnQ9CRDcUXuIkIr+Wnp6O4OBgvP3227DZbJBlGQUFBcjJyQEA7Nu3D6tXr4YkSViyZAnMZjNuuukmpKenw2KxYPHixXA6ndi6dSvWrVuHW2+9FaIo4s4778Qrr7yC0tJSyLKM3bt3w+Fw6PxpichfMKARkV8zGAxYtGgR9u/fj1GjRmHAgAF49tlnUVtbCwAYNWoUVqxYgX79+mHZsmVYsGABTCYTzGYzFi1ahI0bN2LAgAF44YUX8OqrryI5ORkAMHv2bKSkpGDy5Mno378/XnvtNSiKoudHJSI/IqiqqupdBBGRHhYsWICjR4/itdde07sUIqJGeAaNiIiIyMcwoBERERH5GF7iJCIiIvIxPINGRERE5GMY0IiIiIh8DAMaERERkY9hQCMiIiLyMQxoRERERD6GAY2IiIjIx/x/dZzS1Lu7GtIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ef6MuMOfJ7r2"
      },
      "source": [
        "Train and validation accuracy are pretty close and following each other. Thus, we’re not overfitting the data. Also notice that the accuracy keeps increasing as the epoch increases, giving us the intuition that increasing the epoch size is likely to give us a higher accuracy."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t59EnM-oLlHA",
        "outputId": "66e45782-0dfc-4b8b-ccf2-141a9db705f6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 409
        }
      },
      "source": [
        "plt.figure(figsize=(10,6))\n",
        "plt.plot(history.history['loss'], color='black',linewidth=3.0)\n",
        "plt.plot(history.history['val_loss'],color='red',linewidth=2.0)\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q_6CdVTQKCkf"
      },
      "source": [
        "From losses curve, it can be visualised that data is not overfitting as both train and validation loss are trends down closely and the losses go lower with increasing number of epochs."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iyaB_ncnfxcG",
        "outputId": "baa83cce-d204-4dc2-b67e-0b860d44b025",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "nn_pred = nn_model.predict(X_test)\n",
        "nn_pred = (nn_pred>0.5)\n",
        "nn_pred = nn_pred.astype(int)\n",
        "nn_pred[0]\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1, 0, 0, 0])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "19QZOcc4lAC5",
        "outputId": "04cbe686-7428-4626-a2a9-18551b2abaee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        }
      },
      "source": [
        "nn_result = pd.DataFrame(index=Y_test.index, columns=[['HC', 'PD_l', 'PD_m', 'PD_h', 'actual', 'pred']]) \n",
        "nn_result[['HC', 'PD_l', 'PD_m', 'PD_h']] = nn_pred\n",
        "nn_result[['actual']] = Y_test\n",
        "nn_result = nn_result.reset_index()\n",
        "\n",
        "for i in range(len(nn_result)):\n",
        "  if nn_result.loc[i,'HC'][0].astype(int) == 1:\n",
        "    nn_result.loc[i,'pred'] = 'HC'\n",
        "  elif nn_result.loc[i,'PD_l'][0].astype(int) == 1:\n",
        "    nn_result.loc[i,'pred'] = 'PD_l'\n",
        "  elif nn_result.loc[i,'PD_m'][0].astype(int) == 1:\n",
        "    nn_result.loc[i,'pred'] = 'PD_m'\n",
        "  else:\n",
        "    nn_result.loc[i,'pred'] = 'PD_h'\n",
        "\n",
        "\n",
        "for i in range(len(nn_result)):\n",
        "  if nn_result.loc[i,'actual'][0] == 0:\n",
        "    nn_result.loc[i,'actual'] = 'HC'\n",
        "  elif nn_result.loc[i,'actual'][0] == 1:\n",
        "    nn_result.loc[i,'actual'] = 'PD_l'\n",
        "  elif nn_result.loc[i,'actual'][0] == 2:\n",
        "    nn_result.loc[i,'actual'] = 'PD_m'\n",
        "  else:\n",
        "    nn_result.loc[i,'actual'] = 'PD_h'\n",
        "\n",
        "nn_result.head(5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead tr th {\n",
              "        text-align: left;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>PATNO</th>\n",
              "      <th>HC</th>\n",
              "      <th>PD_l</th>\n",
              "      <th>PD_m</th>\n",
              "      <th>PD_h</th>\n",
              "      <th>actual</th>\n",
              "      <th>pred</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3620</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>HC</td>\n",
              "      <td>HC</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>3651</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>HC</td>\n",
              "      <td>HC</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3264</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>HC</td>\n",
              "      <td>HC</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3812</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>HC</td>\n",
              "      <td>HC</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>3169</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>HC</td>\n",
              "      <td>HC</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "  PATNO HC PD_l PD_m PD_h actual pred\n",
              "0  3620  1    0    0    0     HC   HC\n",
              "1  3651  1    0    0    0     HC   HC\n",
              "2  3264  1    0    0    0     HC   HC\n",
              "3  3812  1    0    0    0     HC   HC\n",
              "4  3169  1    0    0    0     HC   HC"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KRe2qrBPO43R"
      },
      "source": [
        "Above table represents comparision of actual value and predicted value of Test set data. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ffjJ5fpQFfU8",
        "outputId": "2ca2b307-7b4d-41f0-8350-5caf27244a52",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "scores.append(round(metrics.accuracy_score(nn_result[['actual']],nn_result[['pred']]),2))\n",
        "classifier_names.append('Neural Net')\n",
        "\n",
        "plt.figure(1, figsize=(6, 4))\n",
        "imp, names = zip(*sorted(zip(scores, classifier_names)))\n",
        "plt.barh(range(len(names)), imp, align = 'center')\n",
        "plt.yticks(range(len(names)), names)\n",
        "plt.xlabel('Classifier performance')\n",
        "plt.ylabel('Classifiers')\n",
        "plt.title('Comparision of different classifiers')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i0aNIU3Wd3xh"
      },
      "source": [
        "Above visualization, illustrates the comparision of models based on its prediction accuracy. It can be infered that Support Vector Machine (SVM) is the best performing model when data is clustered using NMF component vector, with 87% classification accuracy. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FYtxIT65sCCo"
      },
      "source": [
        "##**Evaluation of Best Model**\n",
        "Various evaluation metrics are applied to evalute best performing model, Logistic Regression model\n",
        "1. **Matthews Correlation Coefficient:** \n",
        "\\\n",
        "MCC is calculated from confusion matrix and it represents the correlation between observed and predicted classification. Here, if the value of MCC is 1 then it is perfect prediction if it is -1 then it is total disagreement and if it is 0, it means it is a random guess.\n",
        "\n",
        "2. **Precision:**\n",
        "\\\n",
        "Precision is\n",
        "the ratio of true positive to number of positives in the result. It\n",
        "ranges from 0 – 1.\n",
        "\n",
        "3. **Recall:**\n",
        "\\\n",
        "Recall is the proportion of true positive by all possible\n",
        "relevant data. ie. proportion of actual positives\n",
        "that are correctly identified.\n",
        "\n",
        "4. **F1-Score:**\n",
        "\\\n",
        "It is a harmonic mean of precision and recall and is also\n",
        "known as f1 measure. The value ranges from 0 to 1 and a high\n",
        "value of f1 score represents high classification performance.\n",
        "\n",
        "5. **Confusion Metrix:**\n",
        "\\\n",
        "By using this\n",
        "matrix, we can identify which category is correctly classified\n",
        "and which does not."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ELfSvBNCJX6E",
        "outputId": "474eb9e8-ef23-4add-e60b-f1c6b5f544cc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        }
      },
      "source": [
        "model = SVC(kernel=\"linear\", C=0.025, gamma= 0.0001, class_weight='balanced')\n",
        "model.fit(X_train, Y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SVC(C=0.025, break_ties=False, cache_size=200, class_weight='balanced',\n",
              "    coef0=0.0, decision_function_shape='ovr', degree=3, gamma=0.0001,\n",
              "    kernel='linear', max_iter=-1, probability=False, random_state=None,\n",
              "    shrinking=True, tol=0.001, verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N8dUxxQcSUpm",
        "outputId": "a41d4bb9-84cd-4cbc-d63e-f92debce15d2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 138
        }
      },
      "source": [
        "#BEST PERFORMING MODEL FOR PCA BASED DIMENSIONALITY REDUCTION\n",
        "model = XGBClassifier()\n",
        "model.fit(X_train, Y_train)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
              "              colsample_bynode=1, colsample_bytree=1, gamma=0,\n",
              "              learning_rate=0.1, max_delta_step=0, max_depth=3,\n",
              "              min_child_weight=1, missing=None, n_estimators=100, n_jobs=1,\n",
              "              nthread=None, objective='multi:softprob', random_state=0,\n",
              "              reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
              "              silent=None, subsample=1, verbosity=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o_9NrjXuaBmA",
        "outputId": "a9671423-12e4-4440-c1c1-8ca162ad76f2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 468
        }
      },
      "source": [
        "Y_pred = model.predict(X_test)\n",
        "\n",
        "print(\"Evalution Result for SVM model\\n\") \n",
        "print(\"Matthews Correlation Coefficient:\", round(metrics.matthews_corrcoef(Y_test,Y_pred),3))\n",
        "print('Accuracy:',round(metrics.accuracy_score(Y_test,Y_pred),3))  \n",
        "print('F1 score:',round(metrics.f1_score(Y_test,Y_pred,average='weighted'),3))   \n",
        "print('Recall:',round(metrics.recall_score(Y_test,Y_pred,average='weighted'),3))   \n",
        "print('Precision:',round(metrics.precision_score(Y_test,Y_pred,average='weighted'),3))  \n",
        "print('\\n clasification report:\\n',metrics.classification_report(Y_test,Y_pred))  \n",
        "confusion_mtx = metrics.confusion_matrix(Y_test,Y_pred)\n",
        "print('\\n confussion matrix:\\n',confusion_mtx)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Evalution Result for SVM model\n",
            "\n",
            "Matthews Correlation Coefficient: 0.749\n",
            "Accuracy: 0.822\n",
            "F1 score: 0.816\n",
            "Recall: 0.822\n",
            "Precision: 0.823\n",
            "\n",
            " clasification report:\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00        32\n",
            "           1       0.76      0.90      0.82        31\n",
            "           2       0.50      0.45      0.48        11\n",
            "           3       0.82      0.56      0.67        16\n",
            "\n",
            "    accuracy                           0.82        90\n",
            "   macro avg       0.77      0.73      0.74        90\n",
            "weighted avg       0.82      0.82      0.82        90\n",
            "\n",
            "\n",
            " confussion matrix:\n",
            " [[32  0  0  0]\n",
            " [ 0 28  1  2]\n",
            " [ 0  6  5  0]\n",
            " [ 0  3  4  9]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "peKCB8rbFFvK",
        "outputId": "f98fa1c0-5731-47e7-e5e7-8b9488180055",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        }
      },
      "source": [
        "def plot_confusion_matrix(cm, classes,\n",
        "   normalize=True,\n",
        "   title= 'Confusion matrix',\n",
        "   cmap=plt.cm.Greens):\n",
        " \n",
        "   if normalize:\n",
        "     cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        " \n",
        "   plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "   plt.title(title)\n",
        "   plt.colorbar()\n",
        "   tick_marks = np.arange(len(classes))\n",
        "   plt.xticks(tick_marks, classes, rotation=0)\n",
        "   plt.yticks(tick_marks, classes)\n",
        " \n",
        "   fmt = '.2f' if normalize else 'd'\n",
        "   thresh = cm.max() / 2.\n",
        "   for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "      plt.text(j, i, format(cm[i, j], fmt), horizontalalignment='center', color='white' if cm[i, j] > thresh else 'black')\n",
        " \n",
        "   plt.tight_layout()\n",
        "   plt.ylabel('True label')\n",
        "   plt.xlabel('Predicted label')\n",
        "\n",
        "plt.figure(figsize=(6,4))\n",
        "plot_confusion_matrix(confusion_mtx, ['HC', 'PD_low', 'PD_medium', 'PD_high'])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TFJKB-0mPK_t"
      },
      "source": [
        "##Result Discussion\n",
        "\n",
        "In this research, performance of 7 different models are compared. Based on thorough evaluation of model with optimal selection of myper parameters. Best performing model is SVM based model with classification prediction accuracy of 87%. "
      ]
    }
  ]
}