Dense.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import keras\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.layers import Dense\n",
    "from keras.regularizers import l2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_std = pd.read_csv(\"data/final/1min/BidAsk/BTC_std.csv\")\n",
    "Y = pd.read_csv(\"data/final/1min/BTC_Y_std.csv\")\n",
    "\n",
    "#data_std = pd.read_csv(\"data/final/1min/OrderFlow/BTC_std.csv\")\n",
    "#Y = pd.read_csv(\"data/final/1min/BidAsk/BTC_Y_std.csv\")\n",
    "\n",
    "\n",
    "#data_std = pd.read_csv(\"data/final/1min/HF/BTC_std.csv\")\n",
    "#Y = pd.read_csv(\"data/final/1min/BidAsk/BTC_Y_std.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "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>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17043</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17044</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17045</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17046</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17047</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17048 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       y\n",
       "0      0\n",
       "1      2\n",
       "2      2\n",
       "3      2\n",
       "4      1\n",
       "...   ..\n",
       "17043  1\n",
       "17044  0\n",
       "17045  2\n",
       "17046  1\n",
       "17047  1\n",
       "\n",
       "[17048 rows x 1 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def CreateTestTrain(data_std, Y, window):\n",
    "    r,x = data_std.shape\n",
    "    \n",
    "    dat = np.zeros((r-window, window, x))\n",
    "    \n",
    "\n",
    "    for i in range(r-window):\n",
    "        dat[i,:,:] = data_std.iloc[i:i+window,:].values\n",
    "\n",
    "\n",
    "\n",
    "    train = dat[:14000, :,:]\n",
    "    val = dat[14000:15500, :,:]\n",
    "    test = dat[15500:, :,:]\n",
    "\n",
    "    Y_train = Y[window:14005]\n",
    "    Y_val = Y[14000 + window:15500 + window]\n",
    "    Y_test = Y[15500 + window:]\n",
    "    \n",
    "    return train, val, test, Y_train, Y_val, Y_test\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "unexpected EOF while parsing (<ipython-input-23-c1861c929bdf>, line 14)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-23-c1861c929bdf>\"\u001b[0;36m, line \u001b[0;32m14\u001b[0m\n\u001b[0;31m    keras.layers.\u001b[0m\n\u001b[0m                 ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m unexpected EOF while parsing\n"
     ]
    }
   ],
   "source": [
    "model = keras.models.Sequential([\n",
    "keras.layers.Flatten(input_shape=[5, x]),\n",
    "#keras.layers.Dense(256, activation=\"relu\"),\n",
    "#keras.layers.Dropout(0.2), \n",
    "#keras.layers.Dense(128, activation=\"relu\"),\n",
    "#keras.layers.Dropout(0.15), \n",
    "#keras.layers.Dense(64, activation=\"relu\", kernel_regularizer=l2(0.01), bias_regularizer=l2(0.01)),\n",
    "#keras.layers.Dropout(0.1), \n",
    "#keras.layers.Dense(32, activation=\"relu\", kernel_regularizer=l2(0.01), bias_regularizer=l2(0.01)),\n",
    "#keras.layers.Dropout(0.05), \n",
    "keras.layers.Dense(16, activation=\"relu\"),\n",
    "keras.layers.Dense(8, activation=\"relu\"),\n",
    "keras.layers.Dense(4, activation=\"relu\"),\n",
    "keras.layers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Model_1(train,val,test,Y_train,Y_val, Y_test, EP, window, x):\n",
    "    model = keras.models.Sequential([\n",
    "    keras.layers.Flatten(input_shape=[window, x]),\n",
    "    keras.layers.Dense(4, activation=\"relu\"),\n",
    "    keras.layers.Dense(3, activation=\"softmax\")\n",
    "    ])\n",
    "    \n",
    "    \n",
    "    model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
    "    history = model.fit(train, Y_train, epochs=EP, validation_data = (val,Y_val))\n",
    "    \n",
    "    return history, model\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Model_2(train,val,test,Y_train,Y_val, Y_test, EP, window, x):\n",
    "    model = keras.models.Sequential([\n",
    "    keras.layers.Flatten(input_shape=[window, x]),\n",
    "    keras.layers.Dense(8, activation=\"relu\"),\n",
    "    keras.layers.Dense(4, activation=\"relu\"),\n",
    "    keras.layers.Dense(3, activation=\"softmax\")\n",
    "    ])\n",
    "        \n",
    "    model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
    "    history = model.fit(train, Y_train, epochs=EP, validation_data = (val,Y_val))\n",
    "    \n",
    "    return history, model\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Model_3(train,val,test,Y_train,Y_val, Y_test, EP, window, x):\n",
    "    model = keras.models.Sequential([\n",
    "    keras.layers.Flatten(input_shape=[window, x]),\n",
    "    keras.layers.Dense(8, activation=\"relu\"),\n",
    "    keras.layers.Dense(8, activation=\"relu\"),\n",
    "    keras.layers.Dense(4, activation=\"relu\"),\n",
    "    keras.layers.Dense(3, activation=\"softmax\")\n",
    "    ])\n",
    "        \n",
    "    model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
    "    history = model.fit(train, Y_train, epochs=EP, validation_data = (val,Y_val))\n",
    "    \n",
    "    return history, model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Run_Dense(data_std, Y, name, window = 5, EP = 50):\n",
    "    r,x = data_std.shape\n",
    "    \n",
    "    train, val, test, Y_train, Y_val, Y_test = CreateTestTrain(data_std, Y, window)\n",
    "    \n",
    "    His1, Mod1 = Model_1(train,val,test,Y_train,Y_val, Y_test, EP, window, x)\n",
    "    His2, Mod2 = Model_2(train,val,test,Y_train,Y_val, Y_test, EP, window, x)\n",
    "    His3, Mod3 = Model_3(train,val,test,Y_train,Y_val, Y_test, EP, window, x)\n",
    "    \n",
    "    \n",
    "    return test, Y_test, His1, His2, His3, Mod1, Mod2, Mod3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 1.1062 - accuracy: 0.4621 - val_loss: 1.0998 - val_accuracy: 0.4073\n",
      "Epoch 2/50\n",
      "438/438 [==============================] - 0s 986us/step - loss: 0.9821 - accuracy: 0.5698 - val_loss: 1.0976 - val_accuracy: 0.4080\n",
      "Epoch 3/50\n",
      "438/438 [==============================] - 0s 908us/step - loss: 0.9679 - accuracy: 0.5763 - val_loss: 1.1050 - val_accuracy: 0.4067\n",
      "Epoch 4/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9705 - accuracy: 0.5708 - val_loss: 1.1173 - val_accuracy: 0.4080\n",
      "Epoch 5/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9729 - accuracy: 0.5684 - val_loss: 1.1250 - val_accuracy: 0.4100\n",
      "Epoch 6/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9667 - accuracy: 0.5682 - val_loss: 1.1113 - val_accuracy: 0.4067\n",
      "Epoch 7/50\n",
      "438/438 [==============================] - 0s 993us/step - loss: 0.9604 - accuracy: 0.5753 - val_loss: 1.1035 - val_accuracy: 0.4100\n",
      "Epoch 8/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9641 - accuracy: 0.5715 - val_loss: 1.1104 - val_accuracy: 0.4293\n",
      "Epoch 9/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9584 - accuracy: 0.5810 - val_loss: 1.1060 - val_accuracy: 0.4380\n",
      "Epoch 10/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9593 - accuracy: 0.5790 - val_loss: 1.0805 - val_accuracy: 0.4393\n",
      "Epoch 11/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9502 - accuracy: 0.5838 - val_loss: 1.0788 - val_accuracy: 0.4460\n",
      "Epoch 12/50\n",
      "438/438 [==============================] - 0s 982us/step - loss: 0.9605 - accuracy: 0.5716 - val_loss: 1.0918 - val_accuracy: 0.4453\n",
      "Epoch 13/50\n",
      "438/438 [==============================] - 0s 926us/step - loss: 0.9636 - accuracy: 0.5699 - val_loss: 1.0739 - val_accuracy: 0.4487\n",
      "Epoch 14/50\n",
      "438/438 [==============================] - 0s 913us/step - loss: 0.9567 - accuracy: 0.5754 - val_loss: 1.0740 - val_accuracy: 0.4473\n",
      "Epoch 15/50\n",
      "438/438 [==============================] - 0s 940us/step - loss: 0.9511 - accuracy: 0.5815 - val_loss: 1.0781 - val_accuracy: 0.4533\n",
      "Epoch 16/50\n",
      "438/438 [==============================] - 0s 938us/step - loss: 0.9549 - accuracy: 0.5784 - val_loss: 1.0737 - val_accuracy: 0.4467\n",
      "Epoch 17/50\n",
      "438/438 [==============================] - 0s 910us/step - loss: 0.9578 - accuracy: 0.5729 - val_loss: 1.0947 - val_accuracy: 0.4487\n",
      "Epoch 18/50\n",
      "438/438 [==============================] - 0s 936us/step - loss: 0.9552 - accuracy: 0.5758 - val_loss: 1.0745 - val_accuracy: 0.4460\n",
      "Epoch 19/50\n",
      "438/438 [==============================] - 0s 930us/step - loss: 0.9542 - accuracy: 0.5816 - val_loss: 1.0689 - val_accuracy: 0.4467\n",
      "Epoch 20/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9512 - accuracy: 0.5793 - val_loss: 1.0710 - val_accuracy: 0.4480\n",
      "Epoch 21/50\n",
      "438/438 [==============================] - 0s 985us/step - loss: 0.9438 - accuracy: 0.5869 - val_loss: 1.0704 - val_accuracy: 0.4540\n",
      "Epoch 22/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9462 - accuracy: 0.5832 - val_loss: 1.0832 - val_accuracy: 0.4527\n",
      "Epoch 23/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9586 - accuracy: 0.5772 - val_loss: 1.0765 - val_accuracy: 0.4507\n",
      "Epoch 24/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9488 - accuracy: 0.5830 - val_loss: 1.0757 - val_accuracy: 0.4493\n",
      "Epoch 25/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9516 - accuracy: 0.5813 - val_loss: 1.0765 - val_accuracy: 0.4473\n",
      "Epoch 26/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9470 - accuracy: 0.5808 - val_loss: 1.0761 - val_accuracy: 0.4473\n",
      "Epoch 27/50\n",
      "438/438 [==============================] - 0s 990us/step - loss: 0.9484 - accuracy: 0.5826 - val_loss: 1.0782 - val_accuracy: 0.4493\n",
      "Epoch 28/50\n",
      "438/438 [==============================] - 0s 943us/step - loss: 0.9522 - accuracy: 0.5770 - val_loss: 1.0780 - val_accuracy: 0.4513\n",
      "Epoch 29/50\n",
      "438/438 [==============================] - 0s 989us/step - loss: 0.9518 - accuracy: 0.5789 - val_loss: 1.0768 - val_accuracy: 0.4567\n",
      "Epoch 30/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9554 - accuracy: 0.5748 - val_loss: 1.0764 - val_accuracy: 0.4533\n",
      "Epoch 31/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9531 - accuracy: 0.5769 - val_loss: 1.0867 - val_accuracy: 0.4520\n",
      "Epoch 32/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9495 - accuracy: 0.5827 - val_loss: 1.1156 - val_accuracy: 0.4400\n",
      "Epoch 33/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9631 - accuracy: 0.5663 - val_loss: 1.0772 - val_accuracy: 0.4493\n",
      "Epoch 34/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9551 - accuracy: 0.5772 - val_loss: 1.0893 - val_accuracy: 0.4513\n",
      "Epoch 35/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9507 - accuracy: 0.5764 - val_loss: 1.0948 - val_accuracy: 0.4480\n",
      "Epoch 36/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9548 - accuracy: 0.5733 - val_loss: 1.0901 - val_accuracy: 0.4520\n",
      "Epoch 37/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9482 - accuracy: 0.5788 - val_loss: 1.0823 - val_accuracy: 0.4527\n",
      "Epoch 38/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9519 - accuracy: 0.5773 - val_loss: 1.0930 - val_accuracy: 0.4533\n",
      "Epoch 39/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9400 - accuracy: 0.5894 - val_loss: 1.0754 - val_accuracy: 0.4527\n",
      "Epoch 40/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9523 - accuracy: 0.5721 - val_loss: 1.0906 - val_accuracy: 0.4480\n",
      "Epoch 41/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9549 - accuracy: 0.5757 - val_loss: 1.0930 - val_accuracy: 0.4547\n",
      "Epoch 42/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9539 - accuracy: 0.5744 - val_loss: 1.1035 - val_accuracy: 0.4507\n",
      "Epoch 43/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9452 - accuracy: 0.5837 - val_loss: 1.0806 - val_accuracy: 0.4527\n",
      "Epoch 44/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9431 - accuracy: 0.5807 - val_loss: 1.1143 - val_accuracy: 0.4553\n",
      "Epoch 45/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9545 - accuracy: 0.5754 - val_loss: 1.0901 - val_accuracy: 0.4520\n",
      "Epoch 46/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9452 - accuracy: 0.5805 - val_loss: 1.0882 - val_accuracy: 0.4513\n",
      "Epoch 47/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9477 - accuracy: 0.5801 - val_loss: 1.0913 - val_accuracy: 0.4560\n",
      "Epoch 48/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9465 - accuracy: 0.5775 - val_loss: 1.0895 - val_accuracy: 0.4587\n",
      "Epoch 49/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9413 - accuracy: 0.5853 - val_loss: 1.0917 - val_accuracy: 0.4540\n",
      "Epoch 50/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9502 - accuracy: 0.5767 - val_loss: 1.0914 - val_accuracy: 0.4573\n",
      "Epoch 1/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 1.0661 - accuracy: 0.5414 - val_loss: 1.0912 - val_accuracy: 0.4060\n",
      "Epoch 2/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9720 - accuracy: 0.5718 - val_loss: 1.0929 - val_accuracy: 0.4060\n",
      "Epoch 3/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9691 - accuracy: 0.5702 - val_loss: 1.0991 - val_accuracy: 0.4047\n",
      "Epoch 4/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9605 - accuracy: 0.5752 - val_loss: 1.0949 - val_accuracy: 0.4053\n",
      "Epoch 5/50\n",
      "438/438 [==============================] - 0s 993us/step - loss: 0.9564 - accuracy: 0.5785 - val_loss: 1.0915 - val_accuracy: 0.4053\n",
      "Epoch 6/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9618 - accuracy: 0.5735 - val_loss: 1.0948 - val_accuracy: 0.4033\n",
      "Epoch 7/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9631 - accuracy: 0.5723 - val_loss: 1.0897 - val_accuracy: 0.4047\n",
      "Epoch 8/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9548 - accuracy: 0.5789 - val_loss: 1.0860 - val_accuracy: 0.4033\n",
      "Epoch 9/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9578 - accuracy: 0.5706 - val_loss: 1.0998 - val_accuracy: 0.4053\n",
      "Epoch 10/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9588 - accuracy: 0.5709 - val_loss: 1.0936 - val_accuracy: 0.4067\n",
      "Epoch 11/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9476 - accuracy: 0.5812 - val_loss: 1.0824 - val_accuracy: 0.4033\n",
      "Epoch 12/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9546 - accuracy: 0.5737 - val_loss: 1.0925 - val_accuracy: 0.4033\n",
      "Epoch 13/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9576 - accuracy: 0.5702 - val_loss: 1.0863 - val_accuracy: 0.4047\n",
      "Epoch 14/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9484 - accuracy: 0.5800 - val_loss: 1.0923 - val_accuracy: 0.4047\n",
      "Epoch 15/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9512 - accuracy: 0.5753 - val_loss: 1.0990 - val_accuracy: 0.4053\n",
      "Epoch 16/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9561 - accuracy: 0.5729 - val_loss: 1.0851 - val_accuracy: 0.4087\n",
      "Epoch 17/50\n",
      "438/438 [==============================] - 2s 4ms/step - loss: 0.9563 - accuracy: 0.5713 - val_loss: 1.0965 - val_accuracy: 0.4033\n",
      "Epoch 18/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9421 - accuracy: 0.5805 - val_loss: 1.0886 - val_accuracy: 0.4033\n",
      "Epoch 19/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9500 - accuracy: 0.5769 - val_loss: 1.0924 - val_accuracy: 0.4040\n",
      "Epoch 20/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9485 - accuracy: 0.5758 - val_loss: 1.0886 - val_accuracy: 0.4067\n",
      "Epoch 21/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9447 - accuracy: 0.5780 - val_loss: 1.0899 - val_accuracy: 0.4553\n",
      "Epoch 22/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9481 - accuracy: 0.5800 - val_loss: 1.0872 - val_accuracy: 0.4580\n",
      "Epoch 23/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9520 - accuracy: 0.5764 - val_loss: 1.0971 - val_accuracy: 0.4527\n",
      "Epoch 24/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9506 - accuracy: 0.5764 - val_loss: 1.0923 - val_accuracy: 0.4560\n",
      "Epoch 25/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9389 - accuracy: 0.5879 - val_loss: 1.0939 - val_accuracy: 0.4547\n",
      "Epoch 26/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9411 - accuracy: 0.5849 - val_loss: 1.0891 - val_accuracy: 0.4647\n",
      "Epoch 27/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9435 - accuracy: 0.5816 - val_loss: 1.0896 - val_accuracy: 0.4533\n",
      "Epoch 28/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9329 - accuracy: 0.5906 - val_loss: 1.0852 - val_accuracy: 0.4567\n",
      "Epoch 29/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9403 - accuracy: 0.5841 - val_loss: 1.0954 - val_accuracy: 0.4527\n",
      "Epoch 30/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9306 - accuracy: 0.5915 - val_loss: 1.0916 - val_accuracy: 0.4560\n",
      "Epoch 31/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9443 - accuracy: 0.5824 - val_loss: 1.0918 - val_accuracy: 0.4493\n",
      "Epoch 32/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9386 - accuracy: 0.5889 - val_loss: 1.0988 - val_accuracy: 0.4540\n",
      "Epoch 33/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9357 - accuracy: 0.5873 - val_loss: 1.0927 - val_accuracy: 0.4567\n",
      "Epoch 34/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9345 - accuracy: 0.5890 - val_loss: 1.0927 - val_accuracy: 0.4613\n",
      "Epoch 35/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9454 - accuracy: 0.5825 - val_loss: 1.0991 - val_accuracy: 0.4573\n",
      "Epoch 36/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9344 - accuracy: 0.5880 - val_loss: 1.0948 - val_accuracy: 0.4560\n",
      "Epoch 37/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9285 - accuracy: 0.5886 - val_loss: 1.1013 - val_accuracy: 0.4573\n",
      "Epoch 38/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9364 - accuracy: 0.5869 - val_loss: 1.0887 - val_accuracy: 0.4587\n",
      "Epoch 39/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9425 - accuracy: 0.5785 - val_loss: 1.1055 - val_accuracy: 0.4553\n",
      "Epoch 40/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9291 - accuracy: 0.5939 - val_loss: 1.0960 - val_accuracy: 0.4547\n",
      "Epoch 41/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9335 - accuracy: 0.5854 - val_loss: 1.0919 - val_accuracy: 0.4547\n",
      "Epoch 42/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9319 - accuracy: 0.5926 - val_loss: 1.1054 - val_accuracy: 0.4567\n",
      "Epoch 43/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9359 - accuracy: 0.5833 - val_loss: 1.1115 - val_accuracy: 0.4540\n",
      "Epoch 44/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9218 - accuracy: 0.5982 - val_loss: 1.0989 - val_accuracy: 0.4587\n",
      "Epoch 45/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9350 - accuracy: 0.5882 - val_loss: 1.0966 - val_accuracy: 0.4553\n",
      "Epoch 46/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9311 - accuracy: 0.5905 - val_loss: 1.0986 - val_accuracy: 0.4587\n",
      "Epoch 47/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9382 - accuracy: 0.5840 - val_loss: 1.1028 - val_accuracy: 0.4553: 0s - loss: 0.9385 - accuracy: 0.58\n",
      "Epoch 48/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9254 - accuracy: 0.5929 - val_loss: 1.0956 - val_accuracy: 0.4573\n",
      "Epoch 49/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9266 - accuracy: 0.5937 - val_loss: 1.1041 - val_accuracy: 0.4547\n",
      "Epoch 50/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9192 - accuracy: 0.5947 - val_loss: 1.1002 - val_accuracy: 0.4547\n",
      "Epoch 1/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 1.0579 - accuracy: 0.4644 - val_loss: 1.0894 - val_accuracy: 0.4187\n",
      "Epoch 2/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9643 - accuracy: 0.5744 - val_loss: 1.0716 - val_accuracy: 0.4433\n",
      "Epoch 3/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9627 - accuracy: 0.5741 - val_loss: 1.0598 - val_accuracy: 0.4487\n",
      "Epoch 4/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9628 - accuracy: 0.5746 - val_loss: 1.0643 - val_accuracy: 0.4500\n",
      "Epoch 5/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9609 - accuracy: 0.5777 - val_loss: 1.0726 - val_accuracy: 0.4407\n",
      "Epoch 6/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9519 - accuracy: 0.5818 - val_loss: 1.0660 - val_accuracy: 0.4493\n",
      "Epoch 7/50\n",
      "438/438 [==============================] - 0s 995us/step - loss: 0.9571 - accuracy: 0.5774 - val_loss: 1.0817 - val_accuracy: 0.4567\n",
      "Epoch 8/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9547 - accuracy: 0.5777 - val_loss: 1.0825 - val_accuracy: 0.4407\n",
      "Epoch 9/50\n",
      "438/438 [==============================] - 0s 986us/step - loss: 0.9558 - accuracy: 0.5783 - val_loss: 1.0762 - val_accuracy: 0.4527\n",
      "Epoch 10/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9594 - accuracy: 0.5758 - val_loss: 1.0733 - val_accuracy: 0.4513\n",
      "Epoch 11/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9537 - accuracy: 0.5765 - val_loss: 1.0927 - val_accuracy: 0.4547\n",
      "Epoch 12/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9448 - accuracy: 0.5859 - val_loss: 1.0816 - val_accuracy: 0.4533\n",
      "Epoch 13/50\n",
      "438/438 [==============================] - 0s 985us/step - loss: 0.9487 - accuracy: 0.5829 - val_loss: 1.0807 - val_accuracy: 0.4553\n",
      "Epoch 14/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9497 - accuracy: 0.5792 - val_loss: 1.0852 - val_accuracy: 0.4493\n",
      "Epoch 15/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9512 - accuracy: 0.5811 - val_loss: 1.0856 - val_accuracy: 0.4487\n",
      "Epoch 16/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9497 - accuracy: 0.5801 - val_loss: 1.0881 - val_accuracy: 0.4520\n",
      "Epoch 17/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9507 - accuracy: 0.5782 - val_loss: 1.0939 - val_accuracy: 0.4580\n",
      "Epoch 18/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9541 - accuracy: 0.5792 - val_loss: 1.1061 - val_accuracy: 0.4493\n",
      "Epoch 19/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9514 - accuracy: 0.5809 - val_loss: 1.1103 - val_accuracy: 0.4600\n",
      "Epoch 20/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9449 - accuracy: 0.5864 - val_loss: 1.0976 - val_accuracy: 0.4520\n",
      "Epoch 21/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9463 - accuracy: 0.5848 - val_loss: 1.1079 - val_accuracy: 0.4527\n",
      "Epoch 22/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9496 - accuracy: 0.5820 - val_loss: 1.1043 - val_accuracy: 0.4480\n",
      "Epoch 23/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9399 - accuracy: 0.5876 - val_loss: 1.1059 - val_accuracy: 0.4527\n",
      "Epoch 24/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9408 - accuracy: 0.5870 - val_loss: 1.1166 - val_accuracy: 0.4527\n",
      "Epoch 25/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9358 - accuracy: 0.5914 - val_loss: 1.1032 - val_accuracy: 0.4493\n",
      "Epoch 26/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9407 - accuracy: 0.5900 - val_loss: 1.1180 - val_accuracy: 0.4493\n",
      "Epoch 27/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9366 - accuracy: 0.5892 - val_loss: 1.1293 - val_accuracy: 0.4480\n",
      "Epoch 28/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9361 - accuracy: 0.5915 - val_loss: 1.1336 - val_accuracy: 0.4567\n",
      "Epoch 29/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9285 - accuracy: 0.5981 - val_loss: 1.1426 - val_accuracy: 0.4487\n",
      "Epoch 30/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9381 - accuracy: 0.5892 - val_loss: 1.1480 - val_accuracy: 0.4407\n",
      "Epoch 31/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9337 - accuracy: 0.5932 - val_loss: 1.1435 - val_accuracy: 0.4500\n",
      "Epoch 32/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9344 - accuracy: 0.5956 - val_loss: 1.1335 - val_accuracy: 0.4467\n",
      "Epoch 33/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9350 - accuracy: 0.5933 - val_loss: 1.1398 - val_accuracy: 0.4387\n",
      "Epoch 34/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9375 - accuracy: 0.5893 - val_loss: 1.1493 - val_accuracy: 0.4493\n",
      "Epoch 35/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9331 - accuracy: 0.5962 - val_loss: 1.1283 - val_accuracy: 0.4480\n",
      "Epoch 36/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9378 - accuracy: 0.5855 - val_loss: 1.1427 - val_accuracy: 0.4467\n",
      "Epoch 37/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9396 - accuracy: 0.5889 - val_loss: 1.1350 - val_accuracy: 0.4493\n",
      "Epoch 38/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9281 - accuracy: 0.5938 - val_loss: 1.1545 - val_accuracy: 0.4467\n",
      "Epoch 39/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9263 - accuracy: 0.6002 - val_loss: 1.1375 - val_accuracy: 0.4513\n",
      "Epoch 40/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9356 - accuracy: 0.5944 - val_loss: 1.1498 - val_accuracy: 0.4427\n",
      "Epoch 41/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9263 - accuracy: 0.6009 - val_loss: 1.1504 - val_accuracy: 0.4493\n",
      "Epoch 42/50\n",
      "438/438 [==============================] - 0s 1ms/step - loss: 0.9259 - accuracy: 0.5968 - val_loss: 1.1396 - val_accuracy: 0.4427\n",
      "Epoch 43/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9310 - accuracy: 0.5941 - val_loss: 1.1597 - val_accuracy: 0.4513\n",
      "Epoch 44/50\n",
      "438/438 [==============================] - 1s 1ms/step - loss: 0.9293 - accuracy: 0.5985 - val_loss: 1.1575 - val_accuracy: 0.4447\n",
      "Epoch 45/50\n",
      "438/438 [==============================] - 1s 2ms/step - loss: 0.9247 - accuracy: 0.5976 - val_loss: 1.1565 - val_accuracy: 0.4513\n",
      "Epoch 46/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9320 - accuracy: 0.5935 - val_loss: 1.1791 - val_accuracy: 0.4493\n",
      "Epoch 47/50\n",
      "438/438 [==============================] - 2s 3ms/step - loss: 0.9317 - accuracy: 0.5926 - val_loss: 1.1651 - val_accuracy: 0.4367\n",
      "Epoch 48/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9291 - accuracy: 0.5957 - val_loss: 1.1812 - val_accuracy: 0.4487\n",
      "Epoch 49/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9140 - accuracy: 0.6055 - val_loss: 1.1608 - val_accuracy: 0.4433\n",
      "Epoch 50/50\n",
      "438/438 [==============================] - 1s 3ms/step - loss: 0.9308 - accuracy: 0.5944 - val_loss: 1.1823 - val_accuracy: 0.4447\n"
     ]
    }
   ],
   "source": [
    "test, Y_test, His1, His2, His3, Mod1, Mod2, Mod3 = Run_Dense(data_std, Y, \"test\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0688960552215576"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "min(His1.history[\"val_loss\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "Mods.append(Mod2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.engine.sequential.Sequential at 0x7f97e1604430>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Mods[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<tensorflow.python.keras.engine.sequential.Sequential at 0x7f97e1604430>,\n",
       " <tensorflow.python.keras.engine.sequential.Sequential at 0x7f97e1604280>]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Mods"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.5360693 , 0.24001837, 0.22391239],\n",
       "       [0.52927905, 0.23074456, 0.23997648],\n",
       "       [0.43087754, 0.30636546, 0.262757  ],\n",
       "       ...,\n",
       "       [0.49599108, 0.2139881 , 0.2900208 ],\n",
       "       [0.49516997, 0.21432708, 0.290503  ],\n",
       "       [0.57057583, 0.20854287, 0.22088136]], dtype=float32)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Mod1.predict(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f97e131df10>]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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(His2.history[\"loss\"])\n",
    "plt.plot(His2.history[\"val_loss\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fd264283e80>]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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(His3.history[\"loss\"])\n",
    "plt.plot(His3.history[\"val_loss\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fd263cddf10>]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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(His1.history[\"loss\"])\n",
    "plt.plot(His1.history[\"val_loss\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.37005832793259885"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Accuracy ###\n",
    "pred = np.argmax(Mod1.predict(test), axis=-1)\n",
    "n = pred.shape[0]\n",
    "\n",
    "sum(pred == Y_test.iloc[:,0]) / n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1543,)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False, False, False, ..., False, False, False])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred == 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.37005832793259885"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Accuracy ###\n",
    "pred = np.argmax(Mod2.predict(test), axis=-1)\n",
    "n = pred.shape[0]\n",
    "\n",
    "sum(pred == Y_test.iloc[:,0]) / n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'np' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-47ae038d1d71>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m### Accuracy ###\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mpred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mMod3\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpred\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpred\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mY_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0miloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mn\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'np' is not defined"
     ]
    }
   ],
   "source": [
    "### Accuracy ###\n",
    "pred = np.argmax(Mod3.predict(test), axis=-1)\n",
    "n = pred.shape[0]\n",
    "\n",
    "sum(pred == Y_test.iloc[:,0]) / n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn\n",
    "from sklearn.metrics import roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5570072097420699"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred = Mod1.predict(test)\n",
    "roc_auc_score(pd.get_dummies(Y_test), pred, multi_class=\"ovr\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.545774843339723"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred = Mod2.predict(test)\n",
    "roc_auc_score(pd.get_dummies(Y_test), pred, multi_class=\"ovr\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5451939332823635"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred = Mod3.predict(test)\n",
    "roc_auc_score(pd.get_dummies(Y_test), pred, multi_class=\"ovr\")"
   ]
  },
  {
   "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}