student_admission_tensorflow.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true,
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Student Admission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:47.039467Z",
     "start_time": "2021-02-22T13:03:45.813661Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "np.set_printoptions(precision=3)\n",
    "import tensorflow as tf\n",
    "import seaborn as sb\n",
    "from matplotlib import pyplot as plt\n",
    "# import tensorflow.compat.v1 as tf\n",
    "# tf.disable_v2_behavior()\n",
    "import imblearn\n",
    "from yapf.yapflib.yapf_api import FormatCode\n",
    "from pprint import pformat\n",
    "def pprint(x):\n",
    "    print(FormatCode(pformat(x))[0])\n",
    "%matplotlib inline\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn.preprocessing import minmax_scale\n",
    "data = pd.read_csv('https://stats.idre.ucla.edu/stat/data/binary.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "outputs": [
    {
     "data": {
      "text/plain": "<seaborn.axisgrid.PairGrid at 0x2b98ce89cc8>"
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": "<Figure size 582.375x540 with 12 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sb.pairplot(data, hue='admit', palette='hls', vars=['gre', 'rank', 'gpa'])"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "outputs": [
    {
     "data": {
      "text/plain": "              gre       gpa      rank\nadmit                                \n0      573.186813  3.343700  2.641026\n1      618.897638  3.489213  2.149606",
      "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>gre</th>\n      <th>gpa</th>\n      <th>rank</th>\n    </tr>\n    <tr>\n      <th>admit</th>\n      <th></th>\n      <th></th>\n      <th></th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>573.186813</td>\n      <td>3.343700</td>\n      <td>2.641026</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>618.897638</td>\n      <td>3.489213</td>\n      <td>2.149606</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby('admit').mean()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "outputs": [
    {
     "data": {
      "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x2b9a7c29908>"
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sb.boxplot(x=data.admit, y=data['gpa'])"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:47.071655Z",
     "start_time": "2021-02-22T13:03:47.043481Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(400, 9)\n"
     ]
    }
   ],
   "source": [
    "one_hot_data = pd.concat([data, pd.get_dummies(data['rank'], prefix='rank')], axis=1)\n",
    "one_hot_data = pd.concat([one_hot_data, pd.get_dummies(data['admit'], prefix='admit')], axis=1)\n",
    "\n",
    "# Drop the previous rank column\n",
    "data = one_hot_data.drop('rank', axis=1)\n",
    "data = one_hot_data.drop('admit', axis=1)\n",
    "print(data.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:47.103486Z",
     "start_time": "2021-02-22T13:03:47.075481Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "array([[0.276, 0.776, 0.667, ..., 0.   , 1.   , 0.   ],\n       [0.759, 0.81 , 0.667, ..., 0.   , 1.   , 0.   ],\n       [1.   , 1.   , 0.   , ..., 0.   , 0.   , 0.   ],\n       ...,\n       [0.414, 0.213, 0.333, ..., 1.   , 0.   , 0.   ],\n       [0.828, 0.799, 0.333, ..., 1.   , 0.   , 0.   ],\n       [0.655, 0.937, 0.667, ..., 0.   , 1.   , 0.   ]])"
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = data.iloc[:,[0,1,2,3,4,5,6]]\n",
    "y = data.iloc[:,[7,8]]\n",
    "\n",
    "X = minmax_scale(X, feature_range=(0, 1))\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\program files\\python37\\lib\\site-packages\\sklearn\\utils\\validation.py:72: FutureWarning: Pass classes=[0 1] as keyword args. From version 1.0 (renaming of 0.25) passing these as positional arguments will result in an error\n",
      "  \"will result in an error\", FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "over = imblearn.over_sampling.SMOTE(sampling_strategy=0.8, random_state=42)\n",
    "under = imblearn.under_sampling.RandomUnderSampler(sampling_strategy=0.8, random_state=42)\n",
    "steps = [('o', over), ('u', under)]\n",
    "\n",
    "pipeline = imblearn.pipeline.Pipeline(steps=steps)\n",
    "X_balanced, y_balanced = pipeline.fit_resample(X, y=y.to_numpy())\n",
    "y_balanced = np.eye(2)[y_balanced.flatten()]\n",
    "X_balanced, y_balanced = shuffle(X_balanced, y_balanced, random_state=42)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:47.135479Z",
     "start_time": "2021-02-22T13:03:47.109484Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_balanced,y_balanced, test_size=0.2)\n",
    "X_train.shape, y_train.shape\n",
    "X_train_batches, y_train_batches = np.split(X_train,14),np.split(y_train,14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:47.198487Z",
     "start_time": "2021-02-22T13:03:47.143515Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "n_features = X_train.shape[1]\n",
    "n_labels = y_train.shape[1]\n",
    "\n",
    "features = tf.placeholder(tf.float32, [None, n_features])\n",
    "labels = tf.placeholder(tf.float32, [None, n_labels])\n",
    "\n",
    "w = [\n",
    "    tf.Variable(tf.random_normal((n_features, 256), stddev=(1/tf.sqrt(float(n_features))), seed=42), name='Weights_layer_0'),\n",
    "    tf.Variable(tf.random_normal((256, 64), stddev=(1/tf.sqrt(float(256))), seed=42), name='Weights_layer_1'),\n",
    "    tf.Variable(tf.random_normal((64, 8), stddev=(1/tf.sqrt(float(64))), seed=42), name='Weights_layer_2'),\n",
    "    tf.Variable(tf.random_normal((8, n_labels), stddev=(1/tf.sqrt(float(8))), seed=42), name='Weights_layer_3'),\n",
    "]\n",
    "\n",
    "n_layers = len(w)\n",
    "b = [\n",
    "    tf.Variable(tf.random_normal((256,), stddev=0.01, seed=42), name='Bias_layer_0'),\n",
    "    tf.Variable(tf.random_normal((64,), stddev=0.01, seed=42), name='Bias_layer_1'),\n",
    "    tf.Variable(tf.random_normal((8,), stddev=0.01, seed=42), name='Bias_layer_2'),\n",
    "    tf.Variable(tf.random_normal((n_labels,), stddev=0.01, seed=42), name='Bias_layer_3'),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:47.214478Z",
     "start_time": "2021-02-22T13:03:47.201479Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def neural_network(input, weights, biases):\n",
    "    for i in range(n_layers-1):\n",
    "        layer = tf.add(tf.matmul(input if i==0 else layer, weights[i]),biases[i])\n",
    "        layer = tf.nn.selu(layer)\n",
    "        if(i!=0):\n",
    "            layer = tf.nn.dropout(layer, keep_prob=(0.5 + i/10))\n",
    "    out_layer = tf.add(tf.matmul(layer, weights[-1]),biases[-1])\n",
    "    return out_layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:51.683514Z",
     "start_time": "2021-02-22T13:03:47.219480Z"
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 [0.5510204]\n",
      "1 [0.55867344]\n",
      "2 [0.56122446]\n",
      "3 [0.56122446]\n",
      "4 [0.5790816]\n",
      "5 [0.55867344]\n",
      "6 [0.5688775]\n",
      "7 [0.58673465]\n",
      "8 [0.5229592]\n",
      "9 [0.54846936]\n",
      "10 [0.5229592]\n",
      "11 [0.5969388]\n",
      "12 [0.58928573]\n",
      "13 [0.53316325]\n",
      "14 [0.5663265]\n",
      "15 [0.5637755]\n",
      "16 [0.58673465]\n",
      "17 [0.5561224]\n",
      "18 [0.60459185]\n",
      "19 [0.5637755]\n",
      "20 [0.58418363]\n",
      "21 [0.56122446]\n",
      "22 [0.56122446]\n",
      "23 [0.56122446]\n",
      "24 [0.5408163]\n",
      "25 [0.5969388]\n",
      "26 [0.6505102]\n",
      "27 [0.5688775]\n",
      "28 [0.58673465]\n",
      "29 [0.5688775]\n",
      "30 [0.5561224]\n",
      "31 [0.58418363]\n",
      "32 [0.6147959]\n",
      "33 [0.59438777]\n",
      "34 [0.5688775]\n",
      "35 [0.5153061]\n",
      "36 [0.5535714]\n",
      "37 [0.57142854]\n",
      "38 [0.56122446]\n",
      "39 [0.58418363]\n",
      "40 [0.6020408]\n",
      "41 [0.5790816]\n",
      "42 [0.6020408]\n",
      "43 [0.59183675]\n",
      "44 [0.56122446]\n",
      "45 [0.54846936]\n",
      "46 [0.5969388]\n",
      "47 [0.53061223]\n",
      "48 [0.5688775]\n",
      "49 [0.5969388]\n",
      "50 [0.60459185]\n",
      "51 [0.59183675]\n",
      "52 [0.55867344]\n",
      "53 [0.60459185]\n",
      "54 [0.57142854]\n",
      "55 [0.60459185]\n",
      "56 [0.60459185]\n",
      "57 [0.58418363]\n",
      "58 [0.5969388]\n",
      "59 [0.60714287]\n",
      "60 [0.58418363]\n",
      "61 [0.6096939]\n",
      "62 [0.57142854]\n",
      "63 [0.6020408]\n",
      "64 [0.60459185]\n",
      "65 [0.6020408]\n",
      "66 [0.6096939]\n",
      "67 [0.5688775]\n",
      "68 [0.56122446]\n",
      "69 [0.61989796]\n",
      "70 [0.5969388]\n",
      "71 [0.6096939]\n",
      "72 [0.61989796]\n",
      "73 [0.6122449]\n",
      "74 [0.60459185]\n",
      "75 [0.57142854]\n",
      "76 [0.5663265]\n",
      "77 [0.622449]\n",
      "78 [0.58673465]\n",
      "79 [0.58673465]\n",
      "80 [0.622449]\n",
      "81 [0.60459185]\n",
      "82 [0.59438777]\n",
      "83 [0.60459185]\n",
      "84 [0.6122449]\n",
      "85 [0.6122449]\n",
      "86 [0.58673465]\n",
      "87 [0.61989796]\n",
      "88 [0.5688775]\n",
      "89 [0.6020408]\n",
      "90 [0.60714287]\n",
      "91 [0.61989796]\n",
      "92 [0.5688775]\n",
      "93 [0.625]\n",
      "94 [0.63010204]\n",
      "95 [0.6020408]\n",
      "96 [0.58418363]\n",
      "97 [0.627551]\n",
      "98 [0.622449]\n",
      "99 [0.622449]\n",
      "100 [0.59183675]\n",
      "101 [0.60459185]\n",
      "102 [0.6147959]\n",
      "103 [0.59438777]\n",
      "104 [0.5994898]\n",
      "105 [0.6147959]\n",
      "106 [0.6020408]\n",
      "107 [0.6147959]\n",
      "108 [0.59183675]\n",
      "109 [0.61989796]\n",
      "110 [0.5816326]\n",
      "111 [0.60459185]\n",
      "112 [0.625]\n",
      "113 [0.6020408]\n",
      "114 [0.60714287]\n",
      "115 [0.5969388]\n",
      "116 [0.6147959]\n",
      "117 [0.59183675]\n",
      "118 [0.60714287]\n",
      "119 [0.60714287]\n",
      "120 [0.5663265]\n",
      "121 [0.625]\n",
      "122 [0.6122449]\n",
      "123 [0.6096939]\n",
      "124 [0.59438777]\n",
      "125 [0.5663265]\n",
      "126 [0.6020408]\n",
      "127 [0.61989796]\n",
      "128 [0.5688775]\n",
      "129 [0.58673465]\n",
      "130 [0.5969388]\n",
      "131 [0.5994898]\n",
      "132 [0.6096939]\n",
      "133 [0.6020408]\n",
      "134 [0.6352041]\n",
      "135 [0.622449]\n",
      "136 [0.627551]\n",
      "137 [0.6122449]\n",
      "138 [0.59438777]\n",
      "139 [0.6020408]\n",
      "140 [0.59183675]\n",
      "141 [0.6147959]\n",
      "142 [0.63010204]\n",
      "143 [0.622449]\n",
      "144 [0.6530612]\n",
      "145 [0.65561223]\n",
      "146 [0.5994898]\n",
      "147 [0.63010204]\n",
      "148 [0.6096939]\n",
      "149 [0.64540815]\n",
      "150 [0.622449]\n",
      "151 [0.5969388]\n",
      "152 [0.6147959]\n",
      "153 [0.6377551]\n",
      "154 [0.61734694]\n",
      "155 [0.6122449]\n",
      "156 [0.59183675]\n",
      "157 [0.60459185]\n",
      "158 [0.60714287]\n",
      "159 [0.6096939]\n",
      "160 [0.5969388]\n",
      "161 [0.61734694]\n",
      "162 [0.58418363]\n",
      "163 [0.5969388]\n",
      "164 [0.5994898]\n",
      "165 [0.6122449]\n",
      "166 [0.60459185]\n",
      "167 [0.625]\n",
      "168 [0.61734694]\n",
      "169 [0.5688775]\n",
      "170 [0.60714287]\n",
      "171 [0.5816326]\n",
      "172 [0.6147959]\n",
      "173 [0.60714287]\n",
      "174 [0.5994898]\n",
      "175 [0.622449]\n",
      "176 [0.61989796]\n",
      "177 [0.6096939]\n",
      "178 [0.627551]\n",
      "179 [0.61734694]\n",
      "180 [0.6147959]\n",
      "181 [0.6147959]\n",
      "182 [0.6147959]\n",
      "183 [0.625]\n",
      "184 [0.61734694]\n",
      "185 [0.63265306]\n",
      "186 [0.5765306]\n",
      "187 [0.6122449]\n",
      "188 [0.6096939]\n",
      "189 [0.622449]\n",
      "190 [0.5816326]\n",
      "191 [0.6658163]\n",
      "192 [0.61734694]\n",
      "193 [0.58673465]\n",
      "194 [0.6096939]\n",
      "195 [0.6020408]\n",
      "196 [0.60714287]\n",
      "197 [0.63010204]\n",
      "198 [0.6122449]\n",
      "199 [0.622449]\n",
      "200 [0.6096939]\n",
      "201 [0.63010204]\n",
      "202 [0.6377551]\n",
      "203 [0.61734694]\n",
      "204 [0.59183675]\n",
      "205 [0.63010204]\n",
      "206 [0.59183675]\n",
      "207 [0.5790816]\n",
      "208 [0.5510204]\n",
      "209 [0.627551]\n",
      "210 [0.64285713]\n",
      "211 [0.60714287]\n",
      "212 [0.627551]\n",
      "213 [0.6122449]\n",
      "214 [0.60459185]\n",
      "215 [0.60714287]\n",
      "216 [0.6147959]\n",
      "217 [0.6122449]\n",
      "218 [0.61734694]\n",
      "219 [0.627551]\n",
      "220 [0.6377551]\n",
      "221 [0.64285713]\n",
      "222 [0.58928573]\n",
      "223 [0.61989796]\n",
      "224 [0.627551]\n",
      "225 [0.60459185]\n",
      "226 [0.6352041]\n",
      "227 [0.6096939]\n",
      "228 [0.622449]\n",
      "229 [0.63265306]\n",
      "230 [0.61734694]\n",
      "231 [0.627551]\n",
      "232 [0.61734694]\n",
      "233 [0.59183675]\n",
      "234 [0.63265306]\n",
      "235 [0.5994898]\n",
      "236 [0.63010204]\n",
      "237 [0.60459185]\n",
      "238 [0.59438777]\n",
      "239 [0.63010204]\n",
      "240 [0.6122449]\n",
      "241 [0.6479592]\n",
      "242 [0.5994898]\n",
      "243 [0.6352041]\n",
      "244 [0.625]\n",
      "245 [0.6479592]\n",
      "246 [0.61734694]\n",
      "247 [0.627551]\n",
      "248 [0.622449]\n",
      "249 [0.622449]\n",
      "250 [0.60459185]\n",
      "251 [0.61989796]\n",
      "252 [0.6403061]\n",
      "253 [0.6096939]\n",
      "254 [0.622449]\n",
      "255 [0.61734694]\n",
      "256 [0.622449]\n",
      "257 [0.61734694]\n",
      "258 [0.61734694]\n",
      "259 [0.6122449]\n",
      "260 [0.622449]\n",
      "261 [0.60459185]\n",
      "262 [0.6352041]\n",
      "263 [0.625]\n",
      "264 [0.63010204]\n",
      "265 [0.6096939]\n",
      "266 [0.61989796]\n",
      "267 [0.6530612]\n",
      "268 [0.61989796]\n",
      "269 [0.60714287]\n",
      "270 [0.6020408]\n",
      "271 [0.622449]\n",
      "272 [0.6147959]\n",
      "273 [0.63265306]\n",
      "274 [0.6352041]\n",
      "275 [0.625]\n",
      "276 [0.625]\n",
      "277 [0.61989796]\n",
      "278 [0.58418363]\n",
      "279 [0.6352041]\n",
      "280 [0.6147959]\n",
      "281 [0.67091835]\n",
      "282 [0.6020408]\n",
      "283 [0.64540815]\n",
      "284 [0.65816325]\n",
      "285 [0.64285713]\n",
      "286 [0.622449]\n",
      "287 [0.6147959]\n",
      "288 [0.627551]\n",
      "289 [0.60459185]\n",
      "290 [0.60714287]\n",
      "291 [0.6020408]\n",
      "292 [0.6377551]\n",
      "293 [0.622449]\n",
      "294 [0.6479592]\n",
      "295 [0.6352041]\n",
      "296 [0.6530612]\n",
      "297 [0.6377551]\n",
      "298 [0.6505102]\n",
      "299 [0.63010204]\n",
      "300 [0.627551]\n",
      "301 [0.627551]\n",
      "302 [0.5816326]\n",
      "303 [0.6122449]\n",
      "304 [0.61989796]\n",
      "305 [0.6530612]\n",
      "306 [0.6403061]\n",
      "307 [0.63010204]\n",
      "308 [0.64540815]\n",
      "309 [0.6122449]\n",
      "310 [0.6122449]\n",
      "311 [0.627551]\n",
      "312 [0.63010204]\n",
      "313 [0.63265306]\n",
      "314 [0.63265306]\n",
      "315 [0.60714287]\n",
      "316 [0.627551]\n",
      "317 [0.6352041]\n",
      "318 [0.627551]\n",
      "319 [0.627551]\n",
      "320 [0.6096939]\n",
      "321 [0.6096939]\n",
      "322 [0.65816325]\n",
      "323 [0.63010204]\n",
      "324 [0.63265306]\n",
      "325 [0.60459185]\n",
      "326 [0.60714287]\n",
      "327 [0.61734694]\n",
      "328 [0.61734694]\n",
      "329 [0.6147959]\n",
      "330 [0.6377551]\n",
      "331 [0.63010204]\n",
      "332 [0.622449]\n",
      "333 [0.6505102]\n",
      "334 [0.6147959]\n",
      "335 [0.60459185]\n",
      "336 [0.61734694]\n",
      "337 [0.64540815]\n",
      "338 [0.64540815]\n",
      "339 [0.625]\n",
      "340 [0.63010204]\n",
      "341 [0.59183675]\n",
      "342 [0.6377551]\n",
      "343 [0.6352041]\n",
      "344 [0.6403061]\n",
      "345 [0.6377551]\n",
      "346 [0.6096939]\n",
      "347 [0.6479592]\n",
      "348 [0.622449]\n",
      "349 [0.622449]\n",
      "350 [0.5994898]\n",
      "351 [0.6352041]\n",
      "352 [0.627551]\n",
      "353 [0.6147959]\n",
      "354 [0.61989796]\n",
      "355 [0.6122449]\n",
      "356 [0.6377551]\n",
      "357 [0.61989796]\n",
      "358 [0.63010204]\n",
      "359 [0.6122449]\n",
      "360 [0.6505102]\n",
      "361 [0.6530612]\n",
      "362 [0.64285713]\n",
      "363 [0.622449]\n",
      "364 [0.64285713]\n",
      "365 [0.61989796]\n",
      "366 [0.61989796]\n",
      "367 [0.6352041]\n",
      "368 [0.622449]\n",
      "369 [0.61989796]\n",
      "370 [0.63265306]\n",
      "371 [0.63265306]\n",
      "372 [0.622449]\n",
      "373 [0.627551]\n",
      "374 [0.6352041]\n",
      "375 [0.6505102]\n",
      "376 [0.60714287]\n",
      "377 [0.67091835]\n",
      "378 [0.6122449]\n",
      "379 [0.6403061]\n",
      "380 [0.6147959]\n",
      "381 [0.6530612]\n",
      "382 [0.6352041]\n",
      "383 [0.63265306]\n",
      "384 [0.64285713]\n",
      "385 [0.627551]\n",
      "386 [0.6377551]\n",
      "387 [0.625]\n",
      "388 [0.627551]\n",
      "389 [0.627551]\n",
      "390 [0.61734694]\n",
      "391 [0.6122449]\n",
      "392 [0.6147959]\n",
      "393 [0.61989796]\n",
      "394 [0.622449]\n",
      "395 [0.6505102]\n",
      "396 [0.61734694]\n",
      "397 [0.6530612]\n",
      "398 [0.627551]\n",
      "399 [0.64285713]\n",
      "400 [0.6352041]\n",
      "401 [0.6403061]\n",
      "402 [0.6377551]\n",
      "403 [0.6352041]\n",
      "404 [0.6096939]\n",
      "405 [0.63010204]\n",
      "406 [0.63010204]\n",
      "407 [0.65561223]\n",
      "408 [0.61734694]\n",
      "409 [0.627551]\n",
      "410 [0.627551]\n",
      "411 [0.627551]\n",
      "412 [0.6352041]\n",
      "413 [0.627551]\n",
      "414 [0.63010204]\n",
      "415 [0.622449]\n",
      "416 [0.66071427]\n",
      "417 [0.6632653]\n",
      "418 [0.6403061]\n",
      "419 [0.6352041]\n",
      "420 [0.6403061]\n",
      "421 [0.64285713]\n",
      "422 [0.64285713]\n",
      "423 [0.63010204]\n",
      "424 [0.6505102]\n",
      "425 [0.64285713]\n",
      "426 [0.6377551]\n",
      "427 [0.6377551]\n",
      "428 [0.625]\n",
      "429 [0.6377551]\n",
      "430 [0.61734694]\n",
      "431 [0.625]\n",
      "432 [0.625]\n",
      "433 [0.61989796]\n",
      "434 [0.64285713]\n",
      "435 [0.61989796]\n",
      "436 [0.625]\n",
      "437 [0.6352041]\n",
      "438 [0.6352041]\n",
      "439 [0.64285713]\n",
      "440 [0.6377551]\n",
      "441 [0.6479592]\n",
      "442 [0.65816325]\n",
      "443 [0.625]\n",
      "444 [0.622449]\n",
      "445 [0.6505102]\n",
      "446 [0.64540815]\n",
      "447 [0.6479592]\n",
      "448 [0.6352041]\n",
      "449 [0.622449]\n",
      "450 [0.6377551]\n",
      "451 [0.64285713]\n",
      "452 [0.627551]\n",
      "453 [0.63010204]\n",
      "454 [0.627551]\n",
      "455 [0.6760204]\n",
      "456 [0.64540815]\n",
      "457 [0.622449]\n",
      "458 [0.6403061]\n",
      "459 [0.61989796]\n",
      "460 [0.61989796]\n",
      "461 [0.65561223]\n",
      "462 [0.63010204]\n",
      "463 [0.6377551]\n",
      "464 [0.6377551]\n",
      "465 [0.6377551]\n",
      "466 [0.6147959]\n",
      "467 [0.6505102]\n",
      "468 [0.60714287]\n",
      "469 [0.625]\n",
      "470 [0.6122449]\n",
      "471 [0.6377551]\n",
      "472 [0.63265306]\n",
      "473 [0.6403061]\n",
      "474 [0.6096939]\n",
      "475 [0.63265306]\n",
      "476 [0.6377551]\n",
      "477 [0.6352041]\n",
      "478 [0.622449]\n",
      "479 [0.6096939]\n",
      "480 [0.6403061]\n",
      "481 [0.63265306]\n",
      "482 [0.63010204]\n",
      "483 [0.61989796]\n",
      "484 [0.6403061]\n",
      "485 [0.6377551]\n",
      "486 [0.6352041]\n",
      "487 [0.64285713]\n",
      "488 [0.61734694]\n",
      "489 [0.64540815]\n",
      "490 [0.6403061]\n",
      "491 [0.63265306]\n",
      "492 [0.61734694]\n",
      "493 [0.6377551]\n",
      "494 [0.61734694]\n",
      "495 [0.6352041]\n",
      "496 [0.61989796]\n",
      "497 [0.61989796]\n",
      "498 [0.6479592]\n",
      "499 [0.63010204]\n",
      "0.6632653\n"
     ]
    }
   ],
   "source": [
    "loss_ = []\n",
    "res = []\n",
    "prediction = neural_network(features, w, b)\n",
    "# loss = tf.reduce_mean(-tf.reduce_sum(labels*tf.log(prediction), reduction_indices=1))\n",
    "loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=prediction, labels=labels))\n",
    "optim = tf.train.AdadeltaOptimizer(0.05).minimize(loss)\n",
    "# correct_prediction = tf.equal(tf.round(prediction), labels)\n",
    "correct_prediction = tf.equal(tf.argmax(tf.sigmoid(prediction), 1), tf.argmax(labels, 1))\n",
    "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
    "with tf.device('/dml:0'):\n",
    "    with tf.Session() as sess:\n",
    "        sess.run(tf.global_variables_initializer())\n",
    "        for i in range(500):\n",
    "            for m,n in zip(X_train_batches, y_train_batches):\n",
    "                _, l = sess.run([optim, loss],feed_dict={features: m, labels: n})\n",
    "            loss_.append(l)\n",
    "                # pprint(sess.run(w))\n",
    "                # pprint(sess.run(b))\n",
    "\n",
    "            acc = sess.run([accuracy], feed_dict={features: X_train, labels: y_train})\n",
    "            print(i, acc)\n",
    "        test_accuracy = sess.run(accuracy,feed_dict={features: X_test, labels: y_test})\n",
    "        print(test_accuracy)\n",
    "        res = sess.run(tf.sigmoid(prediction),feed_dict={features: X})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:19:22.755958Z",
     "start_time": "2021-02-22T13:19:22.746924Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": "[<matplotlib.lines.Line2D at 0x2ba32ab9f48>]"
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(loss_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-02-22T13:03:51.951501Z",
     "start_time": "2021-02-22T13:03:51.906503Z"
    },
    "pycharm": {
     "name": "#%%\n"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": "     gre   gpa  rank  rank_1  rank_2  rank_3  rank_4  admit_0  admit_1  pred  \\\n0    380  3.61     3       0       0       1       0        1        0     0   \n1    660  3.67     3       0       0       1       0        0        1     0   \n2    800  4.00     1       1       0       0       0        0        1     1   \n3    640  3.19     4       0       0       0       1        0        1     0   \n4    520  2.93     4       0       0       0       1        1        0     0   \n..   ...   ...   ...     ...     ...     ...     ...      ...      ...   ...   \n395  620  4.00     2       0       1       0       0        1        0     1   \n396  560  3.04     3       0       0       1       0        1        0     0   \n397  460  2.63     2       0       1       0       0        1        0     0   \n398  700  3.65     2       0       1       0       0        1        0     1   \n399  600  3.89     3       0       0       1       0        1        0     0   \n\n     admit     res_0     res_1  \n0        0  0.822865  0.232784  \n1        1  0.509649  0.413209  \n2        1  0.218805  0.706213  \n3        1  0.713674  0.296233  \n4        0  0.803616  0.299519  \n..     ...       ...       ...  \n395      0  0.474836  0.580369  \n396      0  0.792779  0.310936  \n397      0  0.799949  0.398979  \n398      0  0.487431  0.552744  \n399      0  0.690889  0.349329  \n\n[400 rows x 13 columns]",
      "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>gre</th>\n      <th>gpa</th>\n      <th>rank</th>\n      <th>rank_1</th>\n      <th>rank_2</th>\n      <th>rank_3</th>\n      <th>rank_4</th>\n      <th>admit_0</th>\n      <th>admit_1</th>\n      <th>pred</th>\n      <th>admit</th>\n      <th>res_0</th>\n      <th>res_1</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>380</td>\n      <td>3.61</td>\n      <td>3</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0.822865</td>\n      <td>0.232784</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>660</td>\n      <td>3.67</td>\n      <td>3</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0.509649</td>\n      <td>0.413209</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>800</td>\n      <td>4.00</td>\n      <td>1</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>1</td>\n      <td>1</td>\n      <td>0.218805</td>\n      <td>0.706213</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>640</td>\n      <td>3.19</td>\n      <td>4</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0.713674</td>\n      <td>0.296233</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>520</td>\n      <td>2.93</td>\n      <td>4</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0.803616</td>\n      <td>0.299519</td>\n    </tr>\n    <tr>\n      <th>...</th>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n      <td>...</td>\n    </tr>\n    <tr>\n      <th>395</th>\n      <td>620</td>\n      <td>4.00</td>\n      <td>2</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0.474836</td>\n      <td>0.580369</td>\n    </tr>\n    <tr>\n      <th>396</th>\n      <td>560</td>\n      <td>3.04</td>\n      <td>3</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0.792779</td>\n      <td>0.310936</td>\n    </tr>\n    <tr>\n      <th>397</th>\n      <td>460</td>\n      <td>2.63</td>\n      <td>2</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0.799949</td>\n      <td>0.398979</td>\n    </tr>\n    <tr>\n      <th>398</th>\n      <td>700</td>\n      <td>3.65</td>\n      <td>2</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0.487431</td>\n      <td>0.552744</td>\n    </tr>\n    <tr>\n      <th>399</th>\n      <td>600</td>\n      <td>3.89</td>\n      <td>3</td>\n      <td>0</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0.690889</td>\n      <td>0.349329</td>\n    </tr>\n  </tbody>\n</table>\n<p>400 rows × 13 columns</p>\n</div>"
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d1 = data\n",
    "d1['pred'] = np.argmax(res, 1)\n",
    "d1['admit'] = np.argmax(d1.iloc[:,[7,8]].to_numpy(), 1)\n",
    "d1['res_0'] = res[:,[0]]\n",
    "d1['res_1'] = res[:,[1]]\n",
    "d1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "outputs": [
    {
     "data": {
      "text/plain": "array([128, 272], dtype=int64)"
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.bincount(np.equal(np.argmax(res, 1), np.argmax(d1.iloc[:,[7,8]].to_numpy(), 1)))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  }
 ],
 "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"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "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()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}