Initial commit

.gitattributes

@@ -0,0 +1,2 @@
+# Auto detect text files and perform LF normalization
+* text=auto

.ipynb_checkpoints/00-minst-checkpoint.ipynb

@@ -0,0 +1,6 @@
+{
+ "cells": [],
+ "metadata": {},
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

.ipynb_checkpoints/01-KNN-checkpoint.ipynb

@@ -0,0 +1,246 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "start read file\n"
+     ]
+    }
+   ],
+   "source": [
+    "from minst import loadData\n",
+    "dataArr, labelArr=loadData('D:/Jupyter/mnist_train.csv')\n",
+    "testDataArr, testLabelArr = loadData('D:/Jupyter/mnist_test.csv')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "def calcDist(x1, x2):\n",
+    "    '''\n",
+    "    计算两个样本点向量之间的距离\n",
+    "    使用的是欧氏距离，即 样本点每个元素相减的平方  再求和  再开方\n",
+    "    :param x1:向量1\n",
+    "    :param x2:向量2\n",
+    "    :return:向量之间的欧式距离\n",
+    "    '''\n",
+    "    return np.sqrt(np.sum(np.square(x1 - x2)))\n",
+    "\n",
+    "    #马哈顿距离计算公式\n",
+    "    # return np.sum(x1 - x2)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def getClosest(trainDataMat, trainLabelMat, x, topK):\n",
+    "    '''\n",
+    "    预测样本x的标记。\n",
+    "    获取方式通过找到与样本x最近的topK个点，并查看它们的标签。\n",
+    "    查找里面占某类标签最多的那类标签\n",
+    "    （书中3.1 3.2节）\n",
+    "    :param trainDataMat:训练集数据集\n",
+    "    :param trainLabelMat:训练集标签集\n",
+    "    :param x:要预测的样本x\n",
+    "    :param topK:选择参考最邻近样本的数目（样本数目的选择关系到正确率，详看3.2.3 K值的选择）\n",
+    "    :return:预测的标记\n",
+    "    '''\n",
+    "    distList = [0] * len(trainLabelMat)        #   [0]*5=[0 0 0 0 0]\n",
+    "\n",
+    "    for i in range(len(trainDataMat)):         #获取训练集中当前样本的向量\n",
+    "\n",
+    "        x1 = trainDataMat[i]                   #计算向量x与训练集样本x的距离\n",
+    "        curDist = calcDist(x1, x)              #将距离放入对应的列表位置中\n",
+    "        distList[i] = curDist\n",
+    "        \n",
+    "    topKList = np.argsort(np.array(distList))[:topK]        #前K个升序排序   \n",
+    "    '''排序的只是索引index！！！！！'''\n",
+    "    \n",
+    "    labelList = [0] * 10                                   \n",
+    "    for index in topKList:\n",
+    "    '''遍历的是array里的元素!!!!'''\n",
+    "    \n",
+    "        #trainLabelMat[index]：在训练集标签中寻找topK元素索引对应的标记\n",
+    "        #int(trainLabelMat[index])：将标记转换为int（实际上已经是int了，但是不int的话，报错）\n",
+    "        #labelList[int(trainLabelMat[index])]：找到标记在labelList中对应的位置\n",
+    "        #最后加1，表示投了一票\n",
+    "        \n",
+    "        labelList[int(trainLabelMat[index])] += 1\n",
+    "        \n",
+    "    #max(labelList)：找到选票箱中票数最多的票数值\n",
+    "    #labelList.index(max(labelList))：再根据最大值在列表中找到该值对应的索引，等同于预测的标记\n",
+    "    return labelList.index(max(labelList))        "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "start read file\n"
+     ]
+    }
+   ],
+   "source": [
+    "trainDataMat = np.mat(dataArr)\n",
+    "trainLabelMat = np.mat(labelArr).T\n",
+    "testDataMat = np.mat(testDataArr)\n",
+    "testLabelMat = np.mat(testLabelArr).T"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "test 0 predict：7 true：7 \n",
+      "test 1 predict：2 true：2 \n",
+      "test 2 predict：1 true：1 \n",
+      "test 3 predict：0 true：0 \n",
+      "test 4 predict：4 true：4 \n",
+      "test 5 predict：1 true：1 \n",
+      "test 6 predict：4 true：4 \n",
+      "test 7 predict：9 true：9 \n",
+      "test 8 predict：5 true：5 \n",
+      "test 9 predict：9 true：9 \n",
+      "test 10 predict：0 true：0 \n",
+      "test 11 predict：6 true：6 \n",
+      "test 12 predict：9 true：9 \n",
+      "test 13 predict：0 true：0 \n",
+      "test 14 predict：1 true：1 \n",
+      "test 15 predict：5 true：5 \n",
+      "test 16 predict：9 true：9 \n",
+      "test 17 predict：7 true：7 \n",
+      "test 18 predict：3 true：3 \n",
+      "test 19 predict：4 true：4 \n",
+      "test 20 predict：9 true：9 \n",
+      "test 21 predict：6 true：6 \n",
+      "test 22 predict：6 true：6 \n",
+      "test 23 predict：5 true：5 \n",
+      "test 24 predict：4 true：4 \n",
+      "test 25 predict：0 true：0 \n",
+      "test 26 predict：7 true：7 \n",
+      "test 27 predict：4 true：4 \n",
+      "test 28 predict：0 true：0 \n",
+      "test 29 predict：1 true：1 \n",
+      "test 30 predict：3 true：3 \n",
+      "test 31 predict：1 true：1 \n",
+      "test 32 predict：3 true：3 \n",
+      "test 33 predict：4 true：4 \n",
+      "test 34 predict：7 true：7 \n",
+      "test 35 predict：2 true：2 \n",
+      "test 36 predict：7 true：7 \n",
+      "test 37 predict：1 true：1 \n",
+      "test 38 predict：2 true：2 \n",
+      "test 39 predict：1 true：1 \n",
+      "test 40 predict：1 true：1 \n",
+      "test 41 predict：7 true：7 \n",
+      "test 42 predict：4 true：4 \n",
+      "test 43 predict：1 true：2 \n",
+      "test 44 predict：3 true：3 \n",
+      "test 45 predict：5 true：5 \n",
+      "test 46 predict：1 true：1 \n",
+      "test 47 predict：2 true：2 \n",
+      "test 48 predict：4 true：4 \n",
+      "test 49 predict：4 true：4 \n",
+      "accuracy：0 \n"
+     ]
+    }
+   ],
+   "source": [
+    "errorCnt = 0\n",
+    "topK = 25\n",
+    "n = 50\n",
+    "for i in range(n):\n",
+    "    x = testDataMat[i]                                    #读取测试集当前测试样本的向量\n",
+    "    y = getClosest(trainDataMat, trainLabelMat, x, topK)  #获取预测的标记\n",
+    "    print('test %d predict：%d true：%d ' % (i,y,testLabelMat[i]))\n",
+    "    if y != testLabelMat[i]: \n",
+    "        errorCnt += 1                                     #如果预测标记与实际标记不符，错误值计数加1\n",
+    "        \n",
+    "accuracy = 1 - (errorCnt / n)\n",
+    "print(accuracy)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.5"
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "oldHeight": 210,
+   "position": {
+    "height": "40px",
+    "left": "705px",
+    "right": "20px",
+    "top": "133px",
+    "width": "481px"
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "varInspector_section_display": "none",
+   "window_display": true
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}