day710_12classificationkeras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(0)\n",
    "n_pts = 100\n",
    "top_reg = np.array([np.random.normal(10,2,n_pts),np.random.normal(10,2,n_pts)]).T\n",
    "bot_reg = np.array([np.random.normal(5,2,n_pts),np.random.normal(5,2,n_pts)]).T\n",
    "xtr = np.vstack((top_reg,bot_reg))\n",
    "ytr = np.matrix(np.append(np.ones(n_pts),np.zeros(n_pts))).T\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input dim:2\n",
      "output dim:1\n"
     ]
    }
   ],
   "source": [
    "print('input dim',xtr.shape[1],sep=':')\n",
    "print('output dim',ytr.shape[1],sep=':')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(xtr[:n_pts,0],xtr[:n_pts,1])\n",
    "plt.scatter(xtr[n_pts:,0],xtr[n_pts:,1])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For Binary (2 class classification),loss=\"binary_crossentropy\"\n",
    "For more than 2 classes , loss=\"categorical_crossentropy\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 1)                 3         \n",
      "=================================================================\n",
      "Total params: 3\n",
      "Trainable params: 3\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(1,input_dim=2,activation='sigmoid'))\n",
    "model.compile(Adam(lr=0.03),loss='binary_crossentropy',metrics=['accuracy'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "200/200 [==============================] - 0s 628us/step - loss: 0.7320 - accuracy: 0.5550\n",
      "Epoch 2/100\n",
      "200/200 [==============================] - 0s 65us/step - loss: 0.6727 - accuracy: 0.5850\n",
      "Epoch 3/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.6163 - accuracy: 0.6150\n",
      "Epoch 4/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.5750 - accuracy: 0.6300\n",
      "Epoch 5/100\n",
      "200/200 [==============================] - 0s 114us/step - loss: 0.5391 - accuracy: 0.6450\n",
      "Epoch 6/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.5116 - accuracy: 0.7050\n",
      "Epoch 7/100\n",
      "200/200 [==============================] - 0s 105us/step - loss: 0.4891 - accuracy: 0.8100\n",
      "Epoch 8/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.4657 - accuracy: 0.8600\n",
      "Epoch 9/100\n",
      "200/200 [==============================] - 0s 110us/step - loss: 0.4593 - accuracy: 0.7500\n",
      "Epoch 10/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.4501 - accuracy: 0.9250\n",
      "Epoch 11/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.4234 - accuracy: 0.8500\n",
      "Epoch 12/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.4072 - accuracy: 0.8150\n",
      "Epoch 13/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.3912 - accuracy: 0.9300\n",
      "Epoch 14/100\n",
      "200/200 [==============================] - 0s 115us/step - loss: 0.3776 - accuracy: 0.8750\n",
      "Epoch 15/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.3689 - accuracy: 0.9350\n",
      "Epoch 16/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 0.3515 - accuracy: 0.90 - 0s 95us/step - loss: 0.3536 - accuracy: 0.9150\n",
      "Epoch 17/100\n",
      "200/200 [==============================] - 0s 110us/step - loss: 0.3494 - accuracy: 0.9000\n",
      "Epoch 18/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.3382 - accuracy: 0.9050\n",
      "Epoch 19/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.3278 - accuracy: 0.9300\n",
      "Epoch 20/100\n",
      "200/200 [==============================] - 0s 125us/step - loss: 0.3203 - accuracy: 0.9400\n",
      "Epoch 21/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.3122 - accuracy: 0.9200\n",
      "Epoch 22/100\n",
      "200/200 [==============================] - 0s 110us/step - loss: 0.3095 - accuracy: 0.9600\n",
      "Epoch 23/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.2967 - accuracy: 0.9350\n",
      "Epoch 24/100\n",
      "200/200 [==============================] - 0s 100us/step - loss: 0.2924 - accuracy: 0.9300\n",
      "Epoch 25/100\n",
      "200/200 [==============================] - 0s 110us/step - loss: 0.2833 - accuracy: 0.9550\n",
      "Epoch 26/100\n",
      "200/200 [==============================] - 0s 105us/step - loss: 0.2783 - accuracy: 0.9450\n",
      "Epoch 27/100\n",
      "200/200 [==============================] - 0s 105us/step - loss: 0.2693 - accuracy: 0.9650\n",
      "Epoch 28/100\n",
      "200/200 [==============================] - 0s 115us/step - loss: 0.2668 - accuracy: 0.9700\n",
      "Epoch 29/100\n",
      "200/200 [==============================] - 0s 105us/step - loss: 0.2596 - accuracy: 0.9600\n",
      "Epoch 30/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.2570 - accuracy: 0.9600\n",
      "Epoch 31/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.2499 - accuracy: 0.9600\n",
      "Epoch 32/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.2457 - accuracy: 0.9650\n",
      "Epoch 33/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.2425 - accuracy: 0.9650\n",
      "Epoch 34/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.2378 - accuracy: 0.9650\n",
      "Epoch 35/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.2348 - accuracy: 0.9500\n",
      "Epoch 36/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.2278 - accuracy: 0.9600\n",
      "Epoch 37/100\n",
      "200/200 [==============================] - 0s 115us/step - loss: 0.2270 - accuracy: 0.9650\n",
      "Epoch 38/100\n",
      "200/200 [==============================] - 0s 115us/step - loss: 0.2248 - accuracy: 0.9650\n",
      "Epoch 39/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.2192 - accuracy: 0.9700\n",
      "Epoch 40/100\n",
      "200/200 [==============================] - 0s 120us/step - loss: 0.2157 - accuracy: 0.9650\n",
      "Epoch 41/100\n",
      "200/200 [==============================] - 0s 269us/step - loss: 0.2121 - accuracy: 0.9650\n",
      "Epoch 42/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.2116 - accuracy: 0.9650\n",
      "Epoch 43/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.2101 - accuracy: 0.9600\n",
      "Epoch 44/100\n",
      "200/200 [==============================] - 0s 145us/step - loss: 0.2048 - accuracy: 0.9650\n",
      "Epoch 45/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.2074 - accuracy: 0.9500\n",
      "Epoch 46/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.2065 - accuracy: 0.9550\n",
      "Epoch 47/100\n",
      "200/200 [==============================] - 0s 100us/step - loss: 0.1977 - accuracy: 0.9550\n",
      "Epoch 48/100\n",
      "200/200 [==============================] - 0s 150us/step - loss: 0.2054 - accuracy: 0.9400\n",
      "Epoch 49/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1960 - accuracy: 0.9550\n",
      "Epoch 50/100\n",
      "200/200 [==============================] - 0s 100us/step - loss: 0.1894 - accuracy: 0.9650\n",
      "Epoch 51/100\n",
      "200/200 [==============================] - 0s 115us/step - loss: 0.1877 - accuracy: 0.9650\n",
      "Epoch 52/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.1839 - accuracy: 0.9650\n",
      "Epoch 53/100\n",
      "200/200 [==============================] - 0s 125us/step - loss: 0.1816 - accuracy: 0.9700\n",
      "Epoch 54/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.1820 - accuracy: 0.9650\n",
      "Epoch 55/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1785 - accuracy: 0.9650\n",
      "Epoch 56/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1812 - accuracy: 0.9650\n",
      "Epoch 57/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.1729 - accuracy: 0.9700\n",
      "Epoch 58/100\n",
      "200/200 [==============================] - 0s 55us/step - loss: 0.1736 - accuracy: 0.9650\n",
      "Epoch 59/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1716 - accuracy: 0.9700\n",
      "Epoch 60/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.1713 - accuracy: 0.9650\n",
      "Epoch 61/100\n",
      "200/200 [==============================] - 0s 65us/step - loss: 0.1681 - accuracy: 0.9650\n",
      "Epoch 62/100\n",
      "200/200 [==============================] - 0s 65us/step - loss: 0.1664 - accuracy: 0.9650\n",
      "Epoch 63/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.1662 - accuracy: 0.9650\n",
      "Epoch 64/100\n",
      "200/200 [==============================] - 0s 100us/step - loss: 0.1636 - accuracy: 0.9700\n",
      "Epoch 65/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.1612 - accuracy: 0.9650\n",
      "Epoch 66/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1610 - accuracy: 0.9650\n",
      "Epoch 67/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.1590 - accuracy: 0.9650\n",
      "Epoch 68/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1593 - accuracy: 0.9650\n",
      "Epoch 69/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1594 - accuracy: 0.9600\n",
      "Epoch 70/100\n",
      "200/200 [==============================] - 0s 100us/step - loss: 0.1592 - accuracy: 0.9600\n",
      "Epoch 71/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.1574 - accuracy: 0.9600\n",
      "Epoch 72/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.1556 - accuracy: 0.9700\n",
      "Epoch 73/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.1525 - accuracy: 0.9700\n",
      "Epoch 74/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1498 - accuracy: 0.9700\n",
      "Epoch 75/100\n",
      "200/200 [==============================] - 0s 70us/step - loss: 0.1507 - accuracy: 0.9650\n",
      "Epoch 76/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.1473 - accuracy: 0.9650\n",
      "Epoch 77/100\n",
      "200/200 [==============================] - 0s 65us/step - loss: 0.1492 - accuracy: 0.9650\n",
      "Epoch 78/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.1454 - accuracy: 0.9650\n",
      "Epoch 79/100\n",
      "200/200 [==============================] - 0s 50us/step - loss: 0.1457 - accuracy: 0.9650\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 80/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1447 - accuracy: 0.9650\n",
      "Epoch 81/100\n",
      "200/200 [==============================] - 0s 70us/step - loss: 0.1435 - accuracy: 0.9650\n",
      "Epoch 82/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.1418 - accuracy: 0.9650\n",
      "Epoch 83/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1443 - accuracy: 0.9650\n",
      "Epoch 84/100\n",
      "200/200 [==============================] - 0s 110us/step - loss: 0.1413 - accuracy: 0.9650\n",
      "Epoch 85/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1391 - accuracy: 0.9650\n",
      "Epoch 86/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1384 - accuracy: 0.9650\n",
      "Epoch 87/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.1433 - accuracy: 0.9600\n",
      "Epoch 88/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1383 - accuracy: 0.9650\n",
      "Epoch 89/100\n",
      "200/200 [==============================] - 0s 85us/step - loss: 0.1391 - accuracy: 0.9650\n",
      "Epoch 90/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1350 - accuracy: 0.9650\n",
      "Epoch 91/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1347 - accuracy: 0.9650\n",
      "Epoch 92/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1343 - accuracy: 0.9650\n",
      "Epoch 93/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.1356 - accuracy: 0.9700\n",
      "Epoch 94/100\n",
      "200/200 [==============================] - 0s 95us/step - loss: 0.1320 - accuracy: 0.9650\n",
      "Epoch 95/100\n",
      "200/200 [==============================] - 0s 110us/step - loss: 0.1325 - accuracy: 0.9650\n",
      "Epoch 96/100\n",
      "200/200 [==============================] - 0s 105us/step - loss: 0.1306 - accuracy: 0.9650\n",
      "Epoch 97/100\n",
      "200/200 [==============================] - 0s 90us/step - loss: 0.1299 - accuracy: 0.9650\n",
      "Epoch 98/100\n",
      "200/200 [==============================] - 0s 70us/step - loss: 0.1291 - accuracy: 0.9650\n",
      "Epoch 99/100\n",
      "200/200 [==============================] - 0s 75us/step - loss: 0.1284 - accuracy: 0.9700\n",
      "Epoch 100/100\n",
      "200/200 [==============================] - 0s 80us/step - loss: 0.1299 - accuracy: 0.9700\n"
     ]
    }
   ],
   "source": [
    "h = model.fit(xtr,ytr,epochs=100,verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(h.history['loss'],label= 'loss')\n",
    "plt.plot(h.history['accuracy'],label= 'accuracy')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_max ,x_min = xtr[:,0].max()+1, xtr[:,0].min()-1\n",
    "y_max , y_min = xtr[:,1].max()+1, xtr[:,1].min()-1\n",
    "xx,yy = np.meshgrid(np.linspace(x_min,x_max),(np.linspace(y_min,y_max)))\n",
    "grid = np.c_[xx.ravel(),yy.ravel()]\n",
    "pred = model.predict(grid).reshape(xx.shape)\n",
    "plt.contourf(xx,yy,pred)\n",
    "plt.scatter(xtr[:n_pts,0],xtr[:n_pts,1])\n",
    "plt.scatter(xtr[n_pts:,0],xtr[n_pts:,1])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}