HW3-classifiers.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "預測趨勢依Close Price-Open Price是否大於0做評估\n",
    "\n",
    "參考資料:\n",
    "\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html\n",
    "\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html\n",
    "\n",
    "https://keras.io/models/sequential/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn import linear_model, metrics, model_selection\n",
    "import random\n",
    "from keras import models, layers, optimizers, regularizers\n",
    "from sklearn import preprocessing\n",
    "import tensorflow as tf\n",
    "from tqdm import tqdm\n",
    "from sklearn import datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_data = pd.read_csv('test.csv')\n",
    "train_data = pd.read_csv('train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "train_data['Move'] = train_data['Close Price']-train_data['Open Price']\n",
    "train_data.loc[train_data['Move'] >= 0, \"Move\"] = 1\n",
    "train_data.loc[train_data['Move'] <= 0, \"Move\"] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>Date</th>\n",
       "      <th>Open Price</th>\n",
       "      <th>Close Price</th>\n",
       "      <th>High Price</th>\n",
       "      <th>Low Price</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>02-Jan-2009</td>\n",
       "      <td>902.99</td>\n",
       "      <td>931.80</td>\n",
       "      <td>934.73</td>\n",
       "      <td>899.35</td>\n",
       "      <td>4048270080</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>05-Jan-2009</td>\n",
       "      <td>929.17</td>\n",
       "      <td>927.45</td>\n",
       "      <td>936.63</td>\n",
       "      <td>919.53</td>\n",
       "      <td>5413910016</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>06-Jan-2009</td>\n",
       "      <td>931.17</td>\n",
       "      <td>934.70</td>\n",
       "      <td>943.85</td>\n",
       "      <td>927.28</td>\n",
       "      <td>5392620032</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>07-Jan-2009</td>\n",
       "      <td>927.45</td>\n",
       "      <td>906.65</td>\n",
       "      <td>927.45</td>\n",
       "      <td>902.37</td>\n",
       "      <td>4704940032</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>08-Jan-2009</td>\n",
       "      <td>905.73</td>\n",
       "      <td>909.73</td>\n",
       "      <td>910.00</td>\n",
       "      <td>896.81</td>\n",
       "      <td>4991549952</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Date  Open Price  Close Price  High Price  Low Price      Volume  \\\n",
       "0  02-Jan-2009      902.99       931.80      934.73     899.35  4048270080   \n",
       "1  05-Jan-2009      929.17       927.45      936.63     919.53  5413910016   \n",
       "2  06-Jan-2009      931.17       934.70      943.85     927.28  5392620032   \n",
       "3  07-Jan-2009      927.45       906.65      927.45     902.37  4704940032   \n",
       "4  08-Jan-2009      905.73       909.73      910.00     896.81  4991549952   \n",
       "\n",
       "   Move  \n",
       "0   1.0  \n",
       "1   0.0  \n",
       "2   1.0  \n",
       "3   0.0  \n",
       "4   1.0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_data['Move'] = test_data['Close Price']-test_data['Open Price']\n",
    "test_data.loc[test_data['Move'] >= 0, \"Move\"] = 1\n",
    "test_data.loc[test_data['Move'] <= 0, \"Move\"] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>Date</th>\n",
       "      <th>Open Price</th>\n",
       "      <th>Close Price</th>\n",
       "      <th>High Price</th>\n",
       "      <th>Low Price</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Move</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>02-Jan-2018</td>\n",
       "      <td>2683.73</td>\n",
       "      <td>2695.81</td>\n",
       "      <td>2695.89</td>\n",
       "      <td>2682.36</td>\n",
       "      <td>1846463232</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>03-Jan-2018</td>\n",
       "      <td>2697.85</td>\n",
       "      <td>2713.06</td>\n",
       "      <td>2714.37</td>\n",
       "      <td>2697.77</td>\n",
       "      <td>2090595328</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>04-Jan-2018</td>\n",
       "      <td>2719.31</td>\n",
       "      <td>2723.99</td>\n",
       "      <td>2729.29</td>\n",
       "      <td>2719.07</td>\n",
       "      <td>2100767744</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>05-Jan-2018</td>\n",
       "      <td>2731.33</td>\n",
       "      <td>2743.15</td>\n",
       "      <td>2743.45</td>\n",
       "      <td>2727.92</td>\n",
       "      <td>1918869120</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>08-Jan-2018</td>\n",
       "      <td>2742.67</td>\n",
       "      <td>2747.71</td>\n",
       "      <td>2748.51</td>\n",
       "      <td>2737.60</td>\n",
       "      <td>1894823936</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Date  Open Price  Close Price  High Price  Low Price      Volume  \\\n",
       "0  02-Jan-2018     2683.73      2695.81     2695.89    2682.36  1846463232   \n",
       "1  03-Jan-2018     2697.85      2713.06     2714.37    2697.77  2090595328   \n",
       "2  04-Jan-2018     2719.31      2723.99     2729.29    2719.07  2100767744   \n",
       "3  05-Jan-2018     2731.33      2743.15     2743.45    2727.92  1918869120   \n",
       "4  08-Jan-2018     2742.67      2747.71     2748.51    2737.60  1894823936   \n",
       "\n",
       "   Move  \n",
       "0   1.0  \n",
       "1   1.0  \n",
       "2   1.0  \n",
       "3   1.0  \n",
       "4   1.0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date           0\n",
       "Open Price     0\n",
       "Close Price    0\n",
       "High Price     0\n",
       "Low Price      0\n",
       "Volume         0\n",
       "Move           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date           0\n",
       "Open Price     0\n",
       "Close Price    0\n",
       "High Price     0\n",
       "Low Price      0\n",
       "Volume         0\n",
       "Move           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "資料沒有遺漏值，不需做遺漏值填補"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_data = test_data.drop(columns = \"Date\")\n",
    "train_data = train_data.drop(columns = \"Date\")\n",
    "test_data = test_data.drop(columns = \"Close Price\")\n",
    "train_data = train_data.drop(columns = \"Close Price\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "此處Date為不重要資訊故移除\n",
    "\n",
    "Close Price為預測結果的一部分，也做移除"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 2264 entries, 0 to 2263\n",
      "Data columns (total 5 columns):\n",
      "Open Price    2264 non-null float64\n",
      "High Price    2264 non-null float64\n",
      "Low Price     2264 non-null float64\n",
      "Volume        2264 non-null int64\n",
      "Move          2264 non-null float64\n",
      "dtypes: float64(4), int64(1)\n",
      "memory usage: 88.5 KB\n"
     ]
    }
   ],
   "source": [
    "train_data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 252 entries, 0 to 251\n",
      "Data columns (total 5 columns):\n",
      "Open Price    252 non-null float64\n",
      "High Price    252 non-null float64\n",
      "Low Price     252 non-null float64\n",
      "Volume        252 non-null int64\n",
      "Move          252 non-null float64\n",
      "dtypes: float64(4), int64(1)\n",
      "memory usage: 9.9 KB\n"
     ]
    }
   ],
   "source": [
    "test_data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split(train_data, test_data):\n",
    "    x_train = train_data.loc[:, train_data.columns != 'Move']\n",
    "    y_train = train_data['Move']\n",
    "    x_test = test_data.loc[:, test_data.columns != 'Move']\n",
    "    y_test = test_data['Move']\n",
    "    return x_train, y_train, x_test, y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_feature(data, feature_name):\n",
    "    plt.figure(figsize=(10, 3))\n",
    "    plt.scatter(data[feature_name], data['Move'])\n",
    "    plt.xlabel(feature_name)\n",
    "    plt.ylabel('Move')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmQAAADTCAYAAAAvQQ9YAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAHLJJREFUeJzt3X+4XHV94PH3505u9AaQkBJXCOFnAQsiIHeBlv5Aqw2iBRR/EKUL6iPaina1Zh9QiohafaRdq8+6u7IurVoXRKUxIhbdiu2zSpAbfhoQDYjkB5UUCP5IICF89o85k0wmM3NnknvumXvv+/U882TOd77nnM/3M2fO/eT8mInMRJIkSdUZqjoASZKkmc6CTJIkqWIWZJIkSRWzIJMkSaqYBZkkSVLFLMgkSZIqZkEmSZJUMQsySZKkilmQSZIkVWxW1QH0a999982DDz646jAkSZLGtWLFin/PzPnj9ZtyBdnBBx/M2NhY1WFIkiSNKyJ+1ks/T1lKkiRVzIJMkiSpYhZkkiRJFbMgkyRJqpgFmSRJUsVKu8syIq4CXgk8kpkvaPN6AJ8ETgc2Audn5m1lxdOLpbev5Yob72Pdhk3sP3eEJYuO5KzjF1QZ0sCYLrmZLuOYDN1yNV4e+83zJUvv5upbVrM1k1oEJx+6Dw8+uom1GzZRi2BrJgualtNYfuvrL37+fL5x18M8vnELAAEkbOuzz5xhMmHDpi3b2uaODLNl6zP8evPWbfEMBTyTsM+cYZ7aspWNW54BYO7IMEfvvxfLH3h8W6yLT1rI6EHzeN91d23rFwG/c+g8Hnx0E+s2bGJusd4nNm3Z9rw5hgVt8vvBr6/cNg6APWbXeCaTTcU69phdY7g2xBObtjBrCIpmAA5/7h5s3PwM6zZsYmR4aFtc/WqMY+W6X7JhUz2WfeYM84oX7se1t65m89bcnjOgsZaR4SGePVzj8Y3bx9hOc5zttpNLlt7NF5c/ROvcc0eGeeWx+3HTj9bvkN/x3tfhIdia9fe28d59+KxjAHbYphrbTSMHmfV1RsCGjVuYM7vGxs1bd4irsf1df+fDO+TqA3989E7bfrvtt/HvHk3LDmD2rCGeenr79nfZGUd3/RwC28dRxN6I5aj99uL79z+2Le49Ztf4yKuO2TbPug2b2LtpnK3LHG9f0Ji3+X1f0Cau1jG39hlvWb3uh6aayA4flN1ecMTvA78CPt+hIDsdeCf1guwk4JOZedJ4yx0dHc0yvvZi6e1rufi6u9m0ZfuHd2S4xkdffcyUfoMnwnTJzXQZx2Toliugax77zfMlS+/mH5Y/1FNcI8M1zj5hAV9dsXaH5U8Hzfld8pU72bK1nH3zIGveTvrZLnbHuScfyOhB83baZifKcC244jXH7lBE7M66hoeCK157LLDz53B4KCDoa9sJYFYtOs7Tbpnd9gVtl1ELSNjyTOe4eo291/3QIImIFZk5Om6/sgqyIoiDges7FGSfAb6bmVcX0/cBp2bmw92WWVZBdsrHvsPaDZt2al8wd4TvXfSSCV/fVDJdcjNdxjEZuuUK6JrHfvN82MU3dDyC0k63Iy5TXbf8zhSN7aTf7WJX1SJ43t7PLjXnzdt+p89Hv8uDareTKmPoZT80SHotyKr8YtgFwOqm6TVF204FWURcAFwAcOCBB5YSzLoOG1Wn9plkuuRmuoxjMuxKrhqv9Ttvv390p2sxBm6LsD0Hk/U+b80sPe/Ny5+IdQ3CdlJlDL3sh6aiKi/qjzZtbT+BmXllZo5m5uj8+eP++sAu2b+ouHttn0mmS26myzgmQ7dcjZfHfvNci3a7gs767T+VdMvvTNEY/2S9z7WI0nPevPyJWNcgbCdVxtDLfmgqqrIgWwMsbJo+AFhXUSwsWXQkI8O1HdpGhmvbLjScyaZLbqbLOCZDt1yNl8d+87z4pIVt29sZGa6x+KSFOy1/OmjO73Bt+had3TRvJ/1sF7tj8UkL226zE2W4Fjts+7u7ruGh6Pg5HB6KvredKGLstr7W17vtC9ouoxb1a8S69ekx9l73Q1NRlacslwEXRsQ11C/qf2K868fK1LgIcDrdsTFRpktupss4JkMvuer0Wr95btzl1s9dlqMHzZvWd1kCM/4uy8Z2MVl3WQKTcpdl8+djd++ybCynyrss2807WXdZjhfXVFPmXZZXA6cC+wI/Bz4ADANk5v8svvbivwGnUf/aizdl5rhX65d1Ub8kSdJEq/yi/sxcPM7rCbyjrPVLkiRNFX5TvyRJUsUsyCRJkipmQSZJklQxCzJJkqSKWZBJkiRVzIJMkiSpYhZkkiRJFbMgkyRJqpgFmSRJUsUsyCRJkipmQSZJklQxCzJJkqSKWZBJkiRVzIJMkiSpYhZkkiRJFbMgkyRJqpgFmSRJUsUsyCRJkipmQSZJklSxUguyiDgtIu6LiFURcVGb1w+MiJsi4vaIuCsiTi8zHkmSpEFUWkEWETXg08DLgaOAxRFxVEu3S4BrM/N44Bzgv5cVjyRJ0qAq8wjZicCqzHwgMzcD1wBntvRJ4DnF872BdSXGI0mSNJDKLMgWAKubptcUbc0uA86NiDXADcA72y0oIi6IiLGIGFu/fn0ZsUqSJFWmzIIs2rRly/Ri4O8z8wDgdOALEbFTTJl5ZWaOZubo/PnzSwhVkiSpOmUWZGuAhU3TB7DzKcm3ANcCZObNwLOBfUuMSZIkaeCUWZDdChweEYdExGzqF+0va+nzEPCHABHxW9QLMs9JSpKkGaW0giwznwYuBG4E7qV+N+XKiLg8Is4ouv0F8NaIuBO4Gjg/M1tPa0qSJE1rs8pceGbeQP1i/ea2S5ue3wOcUmYMkiRJg85v6pckSaqYBZkkSVLFLMgkSZIqZkEmSZJUMQsySZKkilmQSZIkVcyCTJIkqWIWZJIkSRWzIJMkSaqYBZkkSVLFLMgkSZIqZkEmSZJUMQsySZKkilmQSZIkVcyCTJIkqWIWZJIkSRWzIJMkSaqYBZkkSVLFLMgkSZIqZkEmSZJUsVILsog4LSLui4hVEXFRhz6vi4h7ImJlRPyfMuORJEkaRLPKWnBE1IBPAy8D1gC3RsSyzLynqc/hwMXAKZn5eEQ8t6x4JEmSBlWZR8hOBFZl5gOZuRm4Bjizpc9bgU9n5uMAmflIifFIkiQNpDILsgXA6qbpNUVbsyOAIyLiexGxPCJOa7egiLggIsYiYmz9+vUlhStJklSNMguyaNOWLdOzgMOBU4HFwGcjYu5OM2VemZmjmTk6f/78CQ9UkiSpSmUWZGuAhU3TBwDr2vT5WmZuycyfAvdRL9AkSZJmjDILsluBwyPikIiYDZwDLGvpsxR4MUBE7Ev9FOYDJcYkSZI0cPoqyCJij177ZubTwIXAjcC9wLWZuTIiLo+IM4puNwKPRsQ9wE3Aksx8tJ+YJEmSprrIbL2sq02niN8BPgvsmZkHRsSxwNsy88/KDrDV6Ohojo2NTfZqJUmS+hYRKzJzdLx+vR4h+wSwCHgUIDPvBH5/18OTJElSQ8+nLDNzdUvT1gmORZIkaUbq9Zv6VxenLbO4QP9d1K8LkyRJ0m7q9QjZ24F3UP9i1zXAccW0JEmSdlOvR8giM99YaiSSJEkzVK9HyL4fEd+KiLe0+yZ9SZIk7bqeCrLMPBy4BDgauC0iro+Ic0uNTJIkaYbo5y7LH2Tme4ATgceAz5UWlSRJ0gzSU0EWEc+JiPMi4pvA94GHqRdmkiRJ2k29XtR/J/Xfnbw8M28uMR5JkqQZp9eC7NDMzIjYKyL2zMxflRqVJEnSDNLrNWRHR8TtwA+BeyJiRUS8oMS4JEmSZoxeC7Irgfdk5kGZeSDwF0WbJEmSdlOvBdkemXlTYyIzvwvsUUpEkiRJM0yv15A9EBF/CXyhmD4X+Gk5IUmSJM0svR4hezMwH7gO+Mfi+ZvKCkqSJGkm6ekIWWY+Dryr5FgkSZJmpK4FWUQs6/Z6Zp4xseFIkiTNPOMdIfttYDVwNXALEKVHJEmSNMOMV5A9D3gZsBh4A/AN4OrMXFl2YJIkSTNF14v6M3NrZv5TZp4HnAysAr4bEe/sZeERcVpE3BcRqyLioi79XhMRGRGjfUUvSZI0DYx7UX9EPAt4BfWjZAcDn6J+t+V489WAT1M/wrYGuDUilmXmPS399qJ+w8At/QYvSZI0HYx3Uf/ngBcA3wQ+mJk/7GPZJwKrMvOBYlnXAGcC97T0+xDwceC9fSxbkiRp2hjve8j+BDgC+HPg+xHxi+Lxy4j4xTjzLqB+Q0DDmqJtm4g4HliYmdd3W1BEXBARYxExtn79+nFWK0mSNLV0PUKWmb1+cWw77e7IzG0vRgwBnwDOH29BmXklxW9njo6O5jjdJUmSppTdKbjGswZY2DR9ALCuaXov6qdDvxsRD1K/aWCZF/ZLkqSZpsyC7Fbg8Ig4JCJmA+cA275oNjOfyMx9M/PgzDwYWA6ckZljJcYkSZI0cEoryDLzaeBC4EbgXuDazFwZEZdHhN/wL0mSVOjptyx3VWbeANzQ0nZph76nlhmLJEnSoCrzlKUkSZJ6YEEmSZJUMQsySZKkilmQSZIkVcyCTJIkqWIWZJIkSRWzIJMkSaqYBZkkSVLFLMgkSZIqZkEmSZJUMQsySZKkilmQSZIkVcyCTJIkqWIWZJIkSRWzIJMkSaqYBZkkSVLFLMgkSZIqZkEmSZJUMQsySZKkipVakEXEaRFxX0SsioiL2rz+noi4JyLuioh/joiDyoxHkiRpEJVWkEVEDfg08HLgKGBxRBzV0u12YDQzXwh8Bfh4WfFIkiQNqjKPkJ0IrMrMBzJzM3ANcGZzh8y8KTM3FpPLgQNKjEeSJGkglVmQLQBWN02vKdo6eQvwzXYvRMQFETEWEWPr16+fwBAlSZKqV2ZBFm3asm3HiHOBUeCKdq9n5pWZOZqZo/Pnz5/AECVJkqo3q8RlrwEWNk0fAKxr7RQRLwXeD/xBZj5VYjySJEkDqcwjZLcCh0fEIRExGzgHWNbcISKOBz4DnJGZj5QYiyRJ0sAqrSDLzKeBC4EbgXuBazNzZURcHhFnFN2uAPYEvhwRd0TEsg6LkyRJmrbKPGVJZt4A3NDSdmnT85eWuX5JkqSpwG/qlyRJqpgFmSRJUsUsyCRJkipmQSZJklQxCzJJkqSKWZBJkiRVzIJMkiSpYhZkkiRJFbMgkyRJqpgFmSRJUsUsyCRJkipmQSZJklQxCzJJkqSKWZBJkiRVzIJMkiSpYhZkkiRJFbMgkyRJqpgFmSRJUsUsyCRJkipmQSZJklSxWWUuPCJOAz4J1IDPZubHWl5/FvB54ATgUeD1mflgmTFNBUtvX8sVN97Hug2b2H/uCEsWHclZxy+oOqwJNdljXHr7Wi5btpINm7YAsM+cYT7wx0cPbF77yU8VuZyI9bVbDrBL454zu8bGzVtJoBbB4pMW8uGzjmk7zwe/vpLHN27Z1jZ3ZJjLzti+LVyy9G6uvmU1WzN3mn+P2TU+8qpjdoiz07qX3r6W9//j3fx689Zt8w8PwZ7PHubxjVuoRbA1c9u/C5pycPF1d7FpyzPb5pszPMRfvfqF22Jszd2Lnz+fm360fofpb9z18LZxzh0Z5uj992L5A4+3HVezAObMru0Qd6taBPvuOczPf7m57euza8Eez5rFho1bdnofm/Pbmq/GmPYeGSaCHeYf+9ljfHH5QzSiD6B1JLWAZ3Ln9pHhIc4+4QBu+tF61m7Y1HY8jfegkct2/QAWzB3h4N8Y2SmXC1rGufT2tbzvurvYWLyPEfDGkw7caazt3q8IyGSneBpx7jNnmCe3bN22jXTan42X07OOX9Dx89ztc95pfwrbPxdz5wyTCU9s2tL28z23ZQxDxXvX6XPRuu0356Mxb0NjekGX/Uq7tqr/HkSO8+Hc5QVH1IAfAy8D1gC3Aosz856mPn8GvDAz3x4R5wCvyszXd1vu6Ohojo2NlRLzIFh6+1ouvu5uNm3ZvjMcGa7x0VcfU/nGMlEme4xLb1/Lki/fyZZndtzWh2vBFa85duDy2k9+qsjlRKyv3XKGhwICtmzd/j71M+5W55584A5F2dLb17LkK3fusPzmdV/x2mMZ+9lj/MPyh7rGPhT1Pxqt21OzUw6bx/KfPs7WLn3aGa5F2/ga6/2vrzsOYNyxD5rG+9gpv6ccNo/bHnqi45hqQ9F3LqvQGCfAe669g3YhjzfWXdW6PxvvMzIyXOPsExbw1RVrd/o8d2pvjK3d/rQ2FAxBx8/FcC0gO7/eTfO6+9322+1X2sVS5n4zIlZk5ui4/UosyH4buCwzFxXTFwNk5keb+txY9Lk5ImYB/wbMzy5BTfeC7JSPfaft/8wWzB3hexe9pIKIJt5kj7HT+spc5+7oJz+Dkst+19ftPell2b3MX4vg/o+e3vM8C+aO8G9PPDnuEaQqLZg7AtBz7gbJVMjvRKjyPWr+rPT6GWn3fnRqr3psZa+7rP1mrwVZmacsFwCrm6bXACd16pOZT0fEE8BvAP/e3CkiLgAuADjwwAPLincgrOuwsXVqn4ome4zdljuIee0nP4OSy37X10//fsbdrPUPynjzrNuwaadTXYNmELfXXk2F/E6EKt+j5nXvymdkvPZBGdtUXkc3ZV7UH23aWt/lXvqQmVdm5mhmjs6fP39CghtU+xf/C+i1fSqa7DF2W+4g5rWf/AxKLvtdXz/9+xl3s1rsuHsZb579547sNM+g2X/uyEBus72YCvmdCFW+R83r3ZXPyHjtVY+t7HVX/dkqsyBbAyxsmj4AWNepT3HKcm/gsRJjGnhLFh3JyHBth7aR4dq2ixCng8ke45JFR9avI2gxXIuBzGs/+akilxOxvnbLGR6K+rUdPSy73fytFp+0cIfpJYuO3Gn5zetesujIneZpZyhouz01O+WwedTG6dM2jg7xNda7ZNGRPY190DTex075PeWweV3HtCu5rEJjnEsWHUmnkMcb665q3Z+Nt52MDNdYfNLCtp/nTu2NsbXb/mtD0fVzMVzr/no3zevuN3ft9ivtYhmEv7NlnrK8FTg8Ig4B1gLnAG9o6bMMOA+4GXgN8J1u14/NBI0LCgft7o+JNNljbCx3qtxl2U9+qsrl7q6v03J6XXbr/L3cZdmYp9tdlo0+g36XZWuepspdlq35nc53WQKV3mXZ+hnpdJfl6EHz2n7mOrU3VHWXZWMZ3mXZ78IjTgf+lvrXXlyVmR+JiMuBscxcFhHPBr4AHE/9yNg5mflAt2VO94v6JUnS9DEIF/WTmTcAN7S0Xdr0/EngtWXGIEmSNOj8pn5JkqSKWZBJkiRVzIJMkiSpYhZkkiRJFSv1LssyRMR64GdVxzEJ9qXlFws0ocxv+cxxucxv+cxxuWZKfg/KzHG/1X7KFWQzRUSM9XKbrHaN+S2fOS6X+S2fOS6X+d2RpywlSZIqZkEmSZJUMQuywXVl1QFMc+a3fOa4XOa3fOa4XOa3ideQSZIkVcwjZJIkSRWzIJMkSaqYBdkkiYiFEXFTRNwbESsj4s9bXn9vRGRE7FtMR0R8KiJWRcRdEfGipr7nRcRPisd5kz2WQdUpxxFxWUSsjYg7isfpTfNcXOT4vohY1NR+WtG2KiIuqmI8g6bbNhwR7yzytTIiPt7Ubn571GX7/VLTtvtgRNzRNI/57UOXHB8XEcuLHI9FxIlFu/vhPnTJ77ERcXNE3B0RX4+I5zTN4zbckJk+JuEB7Ae8qHi+F/Bj4KhieiFwI/UvvN23aDsd+CYQwMnALUX7POCB4t99iuf7VD2+QXh0yjFwGfDeNv2PAu4EngUcAtwP1IrH/cChwOyiz1FVj6/qR5f8vhj4v8Cziteea34nLr8tff4GuNT8TmyOgW8BLy/aTwe+2/Tc/fDu5/dW4A+K9jcDHyqeuw03PTxCNkky8+HMvK14/kvgXmBB8fIngP8CNN9hcSbw+axbDsyNiP2ARcC3M/OxzHwc+DZw2mSNY5CNk+N2zgSuycynMvOnwCrgxOKxKjMfyMzNwDVF3xmtS37/FPhYZj5VvPZIMYv57cN4229EBPA64Oqiyfz2qUuOE2gctdkbWFc8dz/chy75PRL416Lbt4Gzi+duw00syCoQEQcDxwO3RMQZwNrMvLOl2wJgddP0mqKtU7uaNOe4aLqwOOVwVUTsU7SZ413Ukt8jgN+LiFsi4l8i4j8W3czvLmqz/QL8HvDzzPxJMW1+d0NLjv8zcEVErAb+Gri46GaOd1FLfn8InFG89FrqZ4XA/O7AgmySRcSewFep7wCeBt4PXNqua5u27NKuQnOOM/MXwP8ADgOOAx6mftoHzPEuaZPfWdRP25wMLAGuLY7mmN9d0Ca/DYvZfnQMzO8ua5PjPwXenZkLgXcD/7vRtc3s5ngcbfL7ZuAdEbGC+qnMzY2ubWafsfm1IJtEETFMfSP9YmZeR71IOAS4MyIeBA4AbouI51H/H8HCptkPoH4YvVO7aJtjMvPnmbk1M58B/hf1w+FgjvvWLr/U83VdcVrnB8Az1H802Pz2qUN+iYhZwKuBLzV1N7+7oEOOzwMaz7+M+4hd1mEf/KPM/KPMPIH6fyruL7qb32ZVX8Q2Ux7UK/7PA3/bpc+DbL+o/xXseDHpD4r2ecBPqR+R2Kd4Pq/q8Q3Co1OOgf2anr+b+jULAEez4wWlD1C/mHRW8fwQtl9QenTV46v60SW/bwcuL54fQf1UQ5jficlv8dppwL+0tJnfCcox9WudTi2e/yGwonjufnhi8tu40WeoeP3NxbTbcNNjFpospwB/AtzddNv6+zLzhg79b6B+h88qYCPwJoDMfCwiPkT9rhWo/yF8rLywp5S2OQYWR8Rx1A95Pwi8DSAzV0bEtcA91E8fvyMztwJExIXU73ytAVdl5srJHMiA6pTfq4CrIuKH1E9FnJf1va357U+3fcQ57Hi60u1313Taht8KfLI4EvkkcEHxmvvh/nTK7+ER8Y5i+jrg78BtuJU/nSRJklQxryGTJEmqmAWZJElSxSzIJEmSKmZBJkmSVDELMkmSpIpZkEkaaBFxQER8LSJ+EhH3R8QnI2J2yes8PyLWR8QdEXFPRLy1Q7/RiPhUmbFImhksyCQNrOInmK4Dlmbm4dS/eHZP4COTsPovZeZxwKnAX0XEf2iJbVZmjmXmuyYhFknTnAWZpEH2EuDJzGx8keRW6r+28OaImFMcyfpaRPxTRNwXER9ozBgR50bED4qjXJ+JiFrR/quI+EhE3BkRy1sLrVaZ+Qj1n3o5KCIui4grI+JbwOcj4tSIuL5Y7p4R8XcRcXfxQ/ZnF+1/FBE3R8RtEfHl4nf+JGkHFmSSBtnRwIrmhqz/WPFDwG8WTScCb6T+4/GvLU4j/hbweuCU4ijX1qIPwB7A8sw8FvhX6t/S3lFEHAocSv3b2gFOAM7MzDe0dP1L4InMPCYzXwh8JyL2BS4BXpqZLwLGgPf0kwBJM4M/nSRpkAX1n7zq1v7tzHwUICKuA36X+s+wnADcWj/ryQjwSNF/M3B98XwF8LIO6359RPwu8BTwtuLncgCWZeamNv1fSv0njgDIzMcj4pXAUcD3inlnAzePM2ZJM5AFmaRBthI4u7khIp4DLKR+GvEEdi7YknrB9rnMvLjNMrfk9t+M20rn/eCXMvPCNu2/7tC/XfEY1AvGxR3mkSTAU5aSBts/A3Mi4j8BFNeB/Q3w95m5sejzsoiYFxEjwFnA94r5XhMRzy3mmxcRB5Uc67eAbQVcROwDLAdOiYjfLNrmRMQRJcchaQqyIJM0sIojWa+ifm3YT4AfA08C72vq9v+ALwB3AF8t7ny8h/q1W9+KiLuAbwP7lRzuh4F9IuKHEXEn8OLMXA+cD1xdxLEceH7JcUiagmL7kXtJmloi4nxgtMOpRUmaMjxCJkmSVDGPkEmSJFXMI2SSJEkVsyCTJEmqmAWZJElSxSzIJEmSKmZBJkmSVLH/DxUPYNkF1hQwAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Open Price\tClose Price\tHigh Price\tLow Price\tVolume\n",
    "plot_feature(train_data, 'Open Price')\n",
    "plot_feature(train_data, 'High Price')\n",
    "plot_feature(train_data, 'Low Price')\n",
    "plot_feature(train_data, 'Volume')\n",
    "plot_feature(test_data, 'Open Price')\n",
    "plot_feature(test_data, 'High Price')\n",
    "plot_feature(test_data, 'Low Price')\n",
    "plot_feature(test_data, 'Volume')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model, x_train, y_train, x_test, y_test):\n",
    "    train_preds = model.predict(x_train)\n",
    "    test_preds = model.predict(x_test)\n",
    "    train_acc = metrics.accuracy_score(y_train, train_preds)\n",
    "    test_acc = metrics.accuracy_score(y_test, test_preds)\n",
    "    print('Train accuracy: %s' % train_acc)\n",
    "    print('Test accuracy: %s' % test_acc)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic Regression Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5123674911660777\n",
      "Test accuracy: 0.4087301587301587\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "x_train, y_train, x_test, y_test = split(train_data, test_data)\n",
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(x_train.astype('int'), y_train.astype('int'))\n",
    "evaluate(clf, x_train.astype('int'), y_train.astype('int'), x_test.astype('int'), y_test.astype('int'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5463780918727915\n",
      "Test accuracy: 0.503968253968254\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(x_train.round(), y_train.round())\n",
    "evaluate(clf, x_train.round(), y_train.round(), x_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Low Price\")\n",
    "z_test = x_test.drop(columns = \"Low Price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5463780918727915\n",
      "Test accuracy: 0.503968253968254\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.round(), y_train.round())\n",
    "evaluate(clf, z_train.round(), y_train.round(), z_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5123674911660777\n",
      "Test accuracy: 0.4087301587301587\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(x_train.astype('int'), y_train.astype('int'))\n",
    "evaluate(clf, x_train.astype('int'), y_train.astype('int'), x_test.astype('int'), y_test.astype('int'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5123674911660777\n",
      "Test accuracy: 0.4087301587301587\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.astype('int'), y_train.astype('int'))\n",
    "evaluate(clf, z_train.astype('int'), y_train.astype('int'), z_test.astype('int'), y_test.astype('int'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"High Price\")\n",
    "z_test = x_test.drop(columns = \"High Price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5463780918727915\n",
      "Test accuracy: 0.503968253968254\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.round(), y_train.round())\n",
    "evaluate(clf, z_train.round(), y_train.round(), z_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Open Price\")\n",
    "z_test = x_test.drop(columns = \"Open Price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5463780918727915\n",
      "Test accuracy: 0.503968253968254\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.round(), y_train.round())\n",
    "evaluate(clf, z_train.round(), y_train.round(), z_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Open Price\")\n",
    "z_test = x_test.drop(columns = \"Open Price\")\n",
    "z_train = z_train.drop(columns = \"Low Price\")\n",
    "z_test = z_test.drop(columns = \"Low Price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.5463780918727915\n",
      "Test accuracy: 0.503968253968254\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.round(), y_train.round())\n",
    "evaluate(clf, z_train.round(), y_train.round(), z_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Volume\")\n",
    "z_test = x_test.drop(columns = \"Volume\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.8507067137809188\n",
      "Test accuracy: 0.8333333333333334\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.round(), y_train.round())\n",
    "evaluate(clf, z_train.round(), y_train.round(), z_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.8507067137809188\n",
      "Test accuracy: 0.8333333333333334\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto',max_iter=106).fit(z_train.round(), y_train.round())\n",
    "evaluate(clf, z_train.round(), y_train.round(), z_test.round(), y_test.round())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.8462897526501767\n",
      "Test accuracy: 0.8214285714285714\n"
     ]
    }
   ],
   "source": [
    "clf = LogisticRegression(solver='lbfgs', multi_class = 'auto').fit(z_train.astype('int'), y_train.astype('int'))\n",
    "evaluate(clf, z_train.astype('int'), y_train.astype('int'), z_test.astype('int'), y_test.astype('int'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "各欄位直接去掉小數的Train accuracy和Test accuracy較四捨五入的低。而Train則較Test高。\n",
    "\n",
    "Train accuracy: 0.5463780918727915\n",
    "\n",
    "Test accuracy: 0.503968253968254\n",
    "\n",
    "嘗試過丟掉各項欄位，accuracy幾乎沒變動，除了丟掉Volume這個欄位，Train accuracy accuracy會上升但是四捨五入的資料結果不收斂。\n",
    "\n",
    "增加迭代次數到106次時四捨五入的資料會收斂，\n",
    "\n",
    "此測試資料最高accuracy為丟掉Volume欄位直接去掉小數位的資料，\n",
    "\n",
    "Train accuracy: 0.8507067137809188\n",
    "\n",
    "Test accuracy: 0.8333333333333334"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(x_train.astype('float'))\n",
    "\n",
    "x_train_std = scaler.transform(x_train.astype('float'))\n",
    "x_test_std = scaler.transform(x_test.astype('float'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Penalty = 0.05, Train accuracy = 54.64 %, Test accuracy = 50.40 %\n"
     ]
    }
   ],
   "source": [
    "penalty = 0.05\n",
    "\n",
    "svm = SVC(C=penalty, kernel=\"linear\").fit(x_train_std.astype('int'), y_train.values.astype('int'))\n",
    "acc_train = accuracy_score(y_train.values.astype('int'), svm.predict(x_train_std.astype('int'))) * 100\n",
    "acc_test = accuracy_score(y_test.values.astype('int'), svm.predict(x_test_std.astype('int'))) * 100\n",
    "print(\"Penalty = %.2f, Train accuracy = %.2f %%, Test accuracy = %.2f %%\" % (penalty, acc_train, acc_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Penalty = 0.05, Train accuracy = 54.64 %, Test accuracy = 50.40 %\n"
     ]
    }
   ],
   "source": [
    "penalty = 0.05\n",
    "\n",
    "svm = SVC(C=penalty, kernel=\"linear\").fit(x_train_std.round(), y_train.values.round())\n",
    "acc_train = accuracy_score(y_train.values.round(), svm.predict(x_train_std.round())) * 100\n",
    "acc_test = accuracy_score(y_test.values.round(), svm.predict(x_test_std.round())) * 100\n",
    "print(\"Penalty = %.2f, Train accuracy = %.2f %%, Test accuracy = %.2f %%\" % (penalty, acc_train, acc_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Low Price\")\n",
    "z_test = x_test.drop(columns = \"Low Price\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler.fit(z_train.astype('float'))\n",
    "z_train_std = scaler.transform(z_train.astype('float'))\n",
    "z_test_std = scaler.transform(z_test.astype('float'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Penalty = 0.05, Train accuracy = 54.64 %, Test accuracy = 50.40 %\n"
     ]
    }
   ],
   "source": [
    "svm = SVC(C=penalty, kernel=\"linear\").fit(z_train_std.astype('int'), y_train.values.astype('int'))\n",
    "acc_train = accuracy_score(y_train.values.astype('int'), svm.predict(z_train_std.astype('int'))) * 100\n",
    "acc_test = accuracy_score(y_test.values.astype('int'), svm.predict(z_test_std.astype('int'))) * 100\n",
    "print(\"Penalty = %.2f, Train accuracy = %.2f %%, Test accuracy = %.2f %%\" % (penalty, acc_train, acc_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"High Price\")\n",
    "z_test = x_test.drop(columns = \"High Price\")\n",
    "scaler.fit(z_train.astype('float'))\n",
    "z_train_std = scaler.transform(z_train.astype('float'))\n",
    "z_test_std = scaler.transform(z_test.astype('float'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Penalty = 0.05, Train accuracy = 54.64 %, Test accuracy = 50.40 %\n"
     ]
    }
   ],
   "source": [
    "svm = SVC(C=penalty, kernel=\"linear\").fit(z_train_std.astype('int'), y_train.values.astype('int'))\n",
    "acc_train = accuracy_score(y_train.values.astype('int'), svm.predict(z_train_std.astype('int'))) * 100\n",
    "acc_test = accuracy_score(y_test.values.astype('int'), svm.predict(z_test_std.astype('int'))) * 100\n",
    "print(\"Penalty = %.2f, Train accuracy = %.2f %%, Test accuracy = %.2f %%\" % (penalty, acc_train, acc_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Open Price\")\n",
    "z_test = x_test.drop(columns = \"Open Price\")\n",
    "scaler.fit(z_train.astype('float'))\n",
    "z_train_std = scaler.transform(z_train.astype('float'))\n",
    "z_test_std = scaler.transform(z_test.astype('float'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Penalty = 0.05, Train accuracy = 54.64 %, Test accuracy = 50.40 %\n"
     ]
    }
   ],
   "source": [
    "svm = SVC(C=penalty, kernel=\"linear\").fit(z_train_std.astype('int'), y_train.values.astype('int'))\n",
    "acc_train = accuracy_score(y_train.values.astype('int'), svm.predict(z_train_std.astype('int'))) * 100\n",
    "acc_test = accuracy_score(y_test.values.astype('int'), svm.predict(z_test_std.astype('int'))) * 100\n",
    "print(\"Penalty = %.2f, Train accuracy = %.2f %%, Test accuracy = %.2f %%\" % (penalty, acc_train, acc_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Volume\")\n",
    "z_test = x_test.drop(columns = \"Volume\")\n",
    "scaler.fit(z_train.astype('float'))\n",
    "z_train_std = scaler.transform(z_train.astype('float'))\n",
    "z_test_std = scaler.transform(z_test.astype('float'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Penalty = 0.05, Train accuracy = 54.64 %, Test accuracy = 50.40 %\n"
     ]
    }
   ],
   "source": [
    "svm = SVC(C=penalty, kernel=\"linear\").fit(z_train_std.astype('int'), y_train.values.astype('int'))\n",
    "acc_train = accuracy_score(y_train.values.astype('int'), svm.predict(z_train_std.astype('int'))) * 100\n",
    "acc_test = accuracy_score(y_test.values.astype('int'), svm.predict(z_test_std.astype('int'))) * 100\n",
    "print(\"Penalty = %.2f, Train accuracy = %.2f %%, Test accuracy = %.2f %%\" % (penalty, acc_train, acc_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "不論丟掉哪個欄位以及資料小數點處理為直接丟棄還是四捨五入結果都相同\n",
    "\n",
    "Train accuracy = 54.64 %, Test accuracy = 50.40 %"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "nx_train = preprocessing.normalize(x_train)\n",
    "nx_test = preprocessing.normalize(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "hidden_units = 20    # how many neurons in the hidden layer\n",
    "activation = 'relu'  # activation function for hidden layer\n",
    "l2 = 0.001           # regularization - how much we penalize large parameter values\n",
    "learning_rate = 0.1  # how big our steps are in gradient descent\n",
    "epochs = 20          # how many epochs to train for\n",
    "batch_size = 100      # how many samples to use for each gradient descent update\n",
    "dim = 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\users\\yj_li\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\tensorflow\\python\\framework\\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    }
   ],
   "source": [
    "# create a sequential model\n",
    "model = models.Sequential()\n",
    "\n",
    "# add the hidden layer\n",
    "model.add(layers.Dense(input_dim=dim,\n",
    "                       units=hidden_units, \n",
    "                       activation=activation))\n",
    "\n",
    "# add the output layer\n",
    "model.add(layers.Dense(input_dim=hidden_units,\n",
    "                       units=1,\n",
    "                       activation='sigmoid'))\n",
    "\n",
    "# define our loss function and optimizer\n",
    "model.compile(loss='binary_crossentropy',\n",
    "              # Adam is a kind of gradient descent\n",
    "              optimizer=optimizers.Adam(lr=learning_rate),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\users\\yj_li\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\tensorflow\\python\\ops\\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Train on 2264 samples, validate on 252 samples\n",
      "Epoch 1/10\n",
      "2264/2264 [==============================] - ETA: 23s - loss: 0.7078 - acc: 0.47 - ETA: 1s - loss: 0.7081 - acc: 0.5183 - 1s 586us/step - loss: 0.7015 - acc: 0.4951 - val_loss: 7.9079 - val_acc: 0.5040\n",
      "Epoch 2/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6925 - acc: 0.520 - ETA: 0s - loss: 0.6920 - acc: 0.530 - ETA: 0s - loss: 0.6924 - acc: 0.527 - ETA: 0s - loss: 0.6902 - acc: 0.540 - 0s 94us/step - loss: 0.6890 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 3/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6707 - acc: 0.620 - ETA: 0s - loss: 0.6882 - acc: 0.555 - ETA: 0s - loss: 0.6883 - acc: 0.555 - 0s 76us/step - loss: 0.6895 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 4/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6949 - acc: 0.500 - ETA: 0s - loss: 0.6907 - acc: 0.544 - 0s 45us/step - loss: 0.6898 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 5/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6794 - acc: 0.590 - ETA: 0s - loss: 0.6886 - acc: 0.553 - ETA: 0s - loss: 0.6896 - acc: 0.546 - ETA: 0s - loss: 0.6888 - acc: 0.549 - 0s 85us/step - loss: 0.6894 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 6/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6716 - acc: 0.630 - ETA: 0s - loss: 0.6945 - acc: 0.490 - ETA: 0s - loss: 0.6927 - acc: 0.515 - 0s 65us/step - loss: 0.6914 - acc: 0.5278 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 7/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6902 - acc: 0.560 - 0s 21us/step - loss: 0.6919 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 8/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6864 - acc: 0.590 - ETA: 0s - loss: 0.6864 - acc: 0.562 - 0s 45us/step - loss: 0.6900 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 9/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6924 - acc: 0.520 - ETA: 0s - loss: 0.6908 - acc: 0.555 - ETA: 0s - loss: 0.6923 - acc: 0.505 - ETA: 0s - loss: 0.6916 - acc: 0.529 - 0s 84us/step - loss: 0.6918 - acc: 0.5287 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "Epoch 10/10\n",
      "2264/2264 [==============================] - ETA: 0s - loss: 0.6883 - acc: 0.550 - ETA: 0s - loss: 0.6898 - acc: 0.546 - 0s 43us/step - loss: 0.6894 - acc: 0.5464 - val_loss: 8.1230 - val_acc: 0.4960\n",
      "2264/2264 [==============================] - ETA:  - ETA:  - ETA:  - 0s 71us/step\n",
      "252/252 [==============================] - ETA:  - 0s 112us/step\n",
      "Training accuracy: 0.5463780920834086\n",
      "Testing accuracy: 0.5039682549143595\n"
     ]
    }
   ],
   "source": [
    "# train the parameters\n",
    "history = model.fit(nx_train, y_train, epochs=10, batch_size=batch_size, validation_data=(x_test,y_test), class_weight=\"balance\")\n",
    "\n",
    "# evaluate accuracy\n",
    "train_acc = model.evaluate(nx_train, y_train, batch_size=32)[1]\n",
    "test_acc = model.evaluate(nx_test, y_test, batch_size=32)[1]\n",
    "print('Training accuracy: %s' % train_acc)\n",
    "print('Testing accuracy: %s' % test_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_and_evaluate(model, x_train, y_train, x_test, y_test, n=20):\n",
    "    train_accs = []\n",
    "    test_accs = []\n",
    "    with tqdm(total=n) as progress_bar:\n",
    "        for _ in range(n):\n",
    "            model.fit(\n",
    "                x_train, \n",
    "                y_train, \n",
    "                epochs=epochs, \n",
    "                batch_size=batch_size, \n",
    "                class_weight=\"balance\", \n",
    "                verbose=False)\n",
    "            train_accs.append(model.evaluate(x_train, y_train, batch_size=32, verbose=False)[1])\n",
    "            test_accs.append(model.evaluate(x_test, y_test, batch_size=32, verbose=False)[1])\n",
    "            progress_bar.update()\n",
    "    print('Avgerage Training Accuracy: %s' % np.average(train_accs))\n",
    "    print('Avgerage Testing Accuracy: %s' % np.average(test_accs))\n",
    "    return train_accs, test_accs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:16<00:00,  1.11it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avgerage Training Accuracy: 0.5417402828908638\n",
      "Avgerage Testing Accuracy: 0.5035714294229237\n"
     ]
    }
   ],
   "source": [
    "_, test_accs = train_and_evaluate(model, nx_train, y_train, nx_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "z_train = x_train.drop(columns = \"Volume\")\n",
    "z_test = x_test.drop(columns = \"Volume\")\n",
    "nx_train = preprocessing.normalize(z_train)\n",
    "nx_test = preprocessing.normalize(z_test)\n",
    "dim = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a sequential model\n",
    "model = models.Sequential()\n",
    "\n",
    "# add the hidden layer\n",
    "model.add(layers.Dense(input_dim=dim,\n",
    "                       units=hidden_units, \n",
    "                       activation=activation))\n",
    "\n",
    "# add the output layer\n",
    "model.add(layers.Dense(input_dim=hidden_units,\n",
    "                       units=1,\n",
    "                       activation='sigmoid'))\n",
    "\n",
    "# define our loss function and optimizer\n",
    "model.compile(loss='binary_crossentropy',\n",
    "              # Adam is a kind of gradient descent\n",
    "              optimizer=optimizers.Adam(lr=learning_rate),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:13<00:00,  1.30it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avgerage Training Accuracy: 0.5463780920834085\n",
      "Avgerage Testing Accuracy: 0.5039682549143596\n"
     ]
    }
   ],
   "source": [
    "_, test_accs = train_and_evaluate(model, nx_train, y_train, nx_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "scrolled": true
   },
   "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": [
    "history = model.fit(\n",
    "    nx_train, y_train, epochs=epochs, batch_size=batch_size, \n",
    "    class_weight=\"balance\", verbose=False)\n",
    "losses = history.history['loss']\n",
    "plt.plot(range(len(losses)), losses, 'r')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "先查看資料型態以及填補遺漏值，做SVM前先把資料做標準化，做Neural Network前，把資料做normal。\n",
    "\n",
    "本次預測結果\n",
    "\n",
    "Logistic Regression\n",
    "\n",
    "Train accuracy: 0.5463780918727915\n",
    "\n",
    "Test accuracy: 0.503968253968254\n",
    "\n",
    "---\n",
    "SVM\n",
    "\n",
    "Train accuracy = 54.64 % \n",
    "\n",
    "Test accuracy = 50.40 %\n",
    "\n",
    "---\n",
    "Neural Network\n",
    "\n",
    "Training accuracy: 0.5463780920834086\n",
    "\n",
    "Testing accuracy: 0.5039682549143595\n",
    "\n",
    "---\n",
    "原始資料三個模型跑出來的預測準確率幾乎一樣，\n",
    "\n",
    "若資料型態不同，則跑出來的結果也會不同。\n",
    "\n",
    "而經過調整後，丟棄Volume欄位(看圖形這類資料較分散)的Logistic Regression的Accuracy有更高的預測率。\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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}