Lab05_107062558.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = pd.read_csv('./data/nba/X_train.csv')\n",
    "y_train = pd.read_csv('./data/nba/y_train.csv')\n",
    "\n",
    "X_test = pd.read_csv('./data/nba/X_test.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(52399, 8)\n",
      "Index(['PERIOD', 'GAME_CLOCK', 'SHOT_CLOCK', 'DRIBBLES', 'TOUCH_TIME',\n",
      "       'SHOT_DIST', 'PTS_TYPE', 'CLOSE_DEF_DIST'],\n",
      "      dtype='object')\n",
      "Index(['FGM'], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_train.columns)\n",
    "X_train_columns = X_train.columns\n",
    "print(y_train.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   PERIOD  GAME_CLOCK  SHOT_CLOCK  DRIBBLES  TOUCH_TIME  SHOT_DIST  PTS_TYPE  \\\n",
      "0       2         557        15.5         0         0.7        2.1         2   \n",
      "1       2         151        20.1         0         0.8        2.7         2   \n",
      "2       3         448        14.2         0         0.8        5.1         2   \n",
      "3       1         279        10.6         0         0.6        1.9         2   \n",
      "\n",
      "   CLOSE_DEF_DIST  \n",
      "0             0.0  \n",
      "1             3.4  \n",
      "2             0.1  \n",
      "3             2.7  \n",
      "   FGM\n",
      "0    1\n",
      "1    1\n",
      "2    0\n",
      "3    1\n"
     ]
    }
   ],
   "source": [
    "print(X_train[:4])\n",
    "print(y_train[:4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def pd2np(df):\n",
    "    return df.as_matrix()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 'SHOT_CLOCK', 'SHOT_DIST', 'CLOSE_DEF_DIST'\n",
    "#  ['SHOT_CLOCK', 'TOUCH_TIME', 'SHOT_DIST', 'CLOSE_DEF_DIST']\n",
    "# X_train = pd2np(X_train.drop(['GAME_CLOCK', 'TOUCH_TIME',\n",
    "#        'SHOT_DIST', 'PTS_TYPE', 'CLOSE_DEF_DIST'], axis=1))\n",
    "X_train = pd2np(X_train.drop(['PERIOD', 'GAME_CLOCK', 'DRIBBLES', 'TOUCH_TIME', 'PTS_TYPE'], axis=1))\n",
    "# X_train = pd2np(X_train)\n",
    "y_train = pd2np(y_train)\n",
    "\n",
    "X_test = pd2np(X_test.drop(['PERIOD', 'GAME_CLOCK', 'DRIBBLES', 'TOUCH_TIME', 'PTS_TYPE'], axis=1))\n",
    "# X_test = pd2np(X_test.drop(['PERIOD', 'GAME_CLOCK','DRIBBLES','PTS_TYPE'], axis=1))\n",
    "# X_test = pd2np(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(52399, 3)\n"
     ]
    }
   ],
   "source": [
    "print(type(X_train))\n",
    "print(X_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import Lasso\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.linear_model import Ridge\n",
    "from pylab import *\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.metrics import r2_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sc = StandardScaler()\n",
    "sc.fit(X_train)\n",
    "X_train_std = sc.transform(X_train)\n",
    "X_test_std = sc.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "[Alpha = 0.00]\n",
      "MSE train: 0.23, test: 0.24\n",
      "\n",
      "[Alpha = 0.01]\n",
      "MSE train: 0.24, test: 0.24\n",
      "\n",
      "[Alpha = 0.10]\n",
      "MSE train: 0.25, test: 0.25\n",
      "\n",
      "[Alpha = 1.00]\n",
      "MSE train: 0.25, test: 0.25\n",
      "\n",
      "[Alpha = 10.00]\n",
      "MSE train: 0.25, test: 0.25\n"
     ]
    }
   ],
   "source": [
    " XX_train, X_valid, yy_train, y_valid = train_test_split(\n",
    "    X_train_std, y_train, test_size=0.3, random_state=2)\n",
    "for a in [0.001, 0.01, 0.1, 1, 10]:\n",
    "    lr_lasso = Lasso(alpha=a)\n",
    "    lr_lasso.fit(XX_train, yy_train)\n",
    "\n",
    "    y_train_pred = lr_lasso.predict(XX_train)\n",
    "    y_test_pred = lr_lasso.predict(X_valid)\n",
    "\n",
    "    print('\\n[Alpha = %.2f]' % a )\n",
    "    print('MSE train: %.2f, test: %.2f' % (\n",
    "                    mean_squared_error(yy_train, y_train_pred),\n",
    "                    mean_squared_error(y_valid, y_test_pred),))                                          "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f008321c4e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Selected attributes: Index(['PERIOD', 'GAME_CLOCK', 'SHOT_CLOCK'], dtype='object')\n"
     ]
    }
   ],
   "source": [
    "var_num = X_train_std.shape[1]\n",
    "\n",
    "lr_lasso = Lasso(alpha = 0.01)\n",
    "lr_lasso.fit(X_train_std, y_train)\n",
    "lr_ridge = Ridge(alpha = 0.1)\n",
    "lr_ridge.fit(X_train_std, y_train)\n",
    "\n",
    "plt.scatter(range(var_num),lr_lasso.coef_, label = 'LASSO', color = 'blue')\n",
    "plt.scatter(range(var_num),lr_ridge.coef_, label = 'Ridge', color = 'green')\n",
    "plt.hlines(y=0, xmin=0, xmax=var_num-1, color='red', linestyle ='--')\n",
    "plt.xlim(0,var_num-1)\n",
    "plt.legend(loc = 'upper right')\n",
    "plt.xlabel('Coefficients index')\n",
    "plt.ylabel('Coefficients')\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "epsilon = 2e-2\n",
    "idxs = np.where(abs(lr_lasso.coef_) > epsilon)\n",
    "print('Selected attributes: {}'.format(X_train_columns[idxs]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import Perceptron\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.6/site-packages/sklearn/utils/validation.py:578: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Perceptron Misclassified samples: 6027\n",
      "Perceptron Accuracy: 0.61660\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.6/site-packages/ipykernel_launcher.py:9: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  if __name__ == '__main__':\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassifier Misclassified samples: 6010\n",
      "RandomForestClassifier Accuracy: 0.61768\n",
      "DecisionTreeClassifier Misclassified samples: 6063\n",
      "DecisionTreeClassifier Accuracy: 0.61431\n",
      "Final Misclassified samples: 5999\n",
      "Final Accuracy: 0.61838\n"
     ]
    }
   ],
   "source": [
    "ppn = Perceptron(max_iter=1000, eta0=1e-3, shuffle = True,random_state=6)\n",
    "ppn.fit(XX_train, yy_train)\n",
    "y_pred = ppn.predict(X_valid)\n",
    "y_pred = np.reshape(y_pred,(y_pred.shape[0],1))\n",
    "print('Perceptron Misclassified samples: %d' % (y_valid != y_pred).sum())\n",
    "print('Perceptron Accuracy: %.5f' % accuracy_score(y_valid, y_pred))\n",
    "\n",
    "clf = RandomForestClassifier(n_estimators=100,max_depth=7,random_state=2,n_jobs=2)\n",
    "clf.fit(XX_train, yy_train)\n",
    "y_pred1 = clf.predict(X_valid)\n",
    "y_pred1 = np.reshape(y_pred1,(y_pred1.shape[0],1))\n",
    "print('RandomForestClassifier Misclassified samples: %d' % (y_valid != y_pred1).sum())\n",
    "print('RandomForestClassifier Accuracy: %.5f' % accuracy_score(y_valid, y_pred1))\n",
    "\n",
    "tree = DecisionTreeClassifier(criterion='entropy', max_depth=5, random_state=1)\n",
    "tree.fit(XX_train, yy_train)\n",
    "y_pred2 = tree.predict(X_valid)\n",
    "y_pred2 = np.reshape(y_pred2,(y_pred2.shape[0],1))\n",
    "print('DecisionTreeClassifier Misclassified samples: %d' % (y_valid != y_pred2).sum())\n",
    "print('DecisionTreeClassifier Accuracy: %.5f' % accuracy_score(y_valid, y_pred2))\n",
    "for i in range(y_valid.shape[0]):\n",
    "    if not y_pred[i][0] == y_pred1[i][0] == y_pred2[i][0] :\n",
    "        if (y_pred[i][0] + y_pred1[i][0] + y_pred2[i][0])/2 < 1:\n",
    "            y_pred[i][0] = 0\n",
    "        else:\n",
    "            y_pred[i][0] = 1 \n",
    "print('Final Misclassified samples: %d' % (y_valid != y_pred).sum())\n",
    "print('Final Accuracy: %.5f' % accuracy_score(y_valid, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = ppn.predict(X_test_std)\n",
    "y_pred = np.reshape(y_pred,(y_pred.shape[0],1))\n",
    "\n",
    "y_pred1 = clf.predict(X_test_std)\n",
    "y_pred1 = np.reshape(y_pred1,(y_pred1.shape[0],1))\n",
    "\n",
    "y_pred2 = tree.predict(X_test_std)\n",
    "y_pred2 = np.reshape(y_pred2,(y_pred2.shape[0],1))\n",
    "for i in range(X_test_std.shape[0]):\n",
    "    if not y_pred[i][0] == y_pred1[i][0] == y_pred2[i][0] :\n",
    "        if (y_pred[i][0] + y_pred1[i][0] + y_pred2[i][0])/2 < 1:\n",
    "            y_pred[i][0] = 0\n",
    "        else:\n",
    "            y_pred[i][0] = 1 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.learning_curve import learning_curve\n",
    "# poly = PolynomialFeatures(degree=d)\n",
    "# X = poly.fit_transform(x[:,np.newaxis])\n",
    "\n",
    "# lr = LinearRegression()\n",
    "# c, r = XX_train.shape\n",
    "# X = XX_train.reshape(c,)\n",
    "c, r = y_train.shape\n",
    "y_train = y_train.reshape(c,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# d = {'col1': [1, 2], 'col2': [3, 4]}\n",
    "d = {}\n",
    "for i in range(X_test_std.shape[0]):\n",
    "    d[i] = y_pred[i][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = pd.DataFrame(data=d, index=['FGM'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>13090</th>\n",
       "      <th>13091</th>\n",
       "      <th>13092</th>\n",
       "      <th>13093</th>\n",
       "      <th>13094</th>\n",
       "      <th>13095</th>\n",
       "      <th>13096</th>\n",
       "      <th>13097</th>\n",
       "      <th>13098</th>\n",
       "      <th>13099</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>FGM</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 13100 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     0      1      2      3      4      5      6      7      8      9      \\\n",
       "FGM      0      0      0      0      1      1      0      0      0      0   \n",
       "\n",
       "     ...    13090  13091  13092  13093  13094  13095  13096  13097  13098  \\\n",
       "FGM  ...        0      1      0      0      0      0      0      0      0   \n",
       "\n",
       "     13099  \n",
       "FGM      0  \n",
       "\n",
       "[1 rows x 13100 columns]"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df.to_csv('y_pred.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "1 How you evaluate your model\n",
    "A1: the higher accuracy_score the better model\n",
    "2 All models you have tried and the result\n",
    "A2: Perceptron : I tried different random_states and the results varies a lot. So, I ran an experiment \n",
    "        to find better result from random_state 1 to 300 and all results' accuracy range is between 0.4 to 0.61\n",
    "    RandomForestClassifier : I tried different degree and random_state. Different degrees varies but when it achieves\n",
    "        higher degree, the accuracy merely improves slightly. Different random_state varies slightly different so I just\n",
    "        picked one slightly better. The accuracies' range of different degree is between 0.61 to 0.62\n",
    "    DecisionTreeClassifier : I tried different degree and random_state. Different degrees varies but when it achieves\n",
    "        higher degree, the accuracy merely improves slightly. Using different random_states does't make much difference \n",
    "        in accuracy. The accuracies' range of different degree is between 0.61 to 0.62\n",
    "3 Plot the learning curve of your best model and tell if it is over-fit or not\n",
    "A3: the curves are below and they seems not over-fit\n",
    "4 The top-3 features you find and how you find it\n",
    "A4:\n",
    "    I find a better alpha to train lasso and then I train a lasso model to get its coefficients. \n",
    "    According to the coefficients, we can get top-3 features.\n",
    "5 A brief report what you do in this assignment\n",
    "A5: training step\n",
    "    step1 --> train the best Perceptron by tuning random_state, RandomForestClassifier by tuning degree and \n",
    "    random_state and DecisionTreeClassifier by tuning degree\n",
    "    step2 --> use these three classifiers to vote for the final prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def mse(model, X, y):\n",
    "    return accuracy_score(y, model.predict(X))\n",
    "#     return (y != model.predict(X)).sum()\n",
    "#     return ((model.predict(X) - y)**2).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrqQJSimllCtpglJKKeVK/x8g3LtsdUXXAAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f0077469160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_sizes, train_scores, test_scores = learning_curve(estimator=ppn, X=X_train_std, y=y_train, scoring=mse)\n",
    "\n",
    "train_mean = np.mean(train_scores, axis=1)\n",
    "train_std = np.std(train_scores, axis=1)\n",
    "test_mean = np.mean(test_scores, axis=1)\n",
    "test_std = np.std(test_scores, axis=1)\n",
    "\n",
    "plt.plot(train_sizes, train_mean,\n",
    "    color='blue', marker='o',\n",
    "    markersize=2,\n",
    "    label='Training error')\n",
    "plt.fill_between(train_sizes,\n",
    "    train_mean+train_std,\n",
    "    train_mean-train_std,\n",
    "    alpha=0.15, color='blue')\n",
    "\n",
    "plt.plot(train_sizes, test_mean,\n",
    "    color='green', linestyle='--',\n",
    "    marker='s', markersize=2,\n",
    "    label='Testing error')\n",
    "plt.fill_between(train_sizes,\n",
    "    test_mean+test_std,\n",
    "    test_mean-test_std,\n",
    "    alpha=0.15, color='green')\n",
    "\n",
    "plt.hlines(y=0.525 ,xmin=0, xmax=35000, color='red', linewidth=2, linestyle='--')\n",
    "plt.title('Perceptron')\n",
    "plt.grid()\n",
    "plt.xlabel('Number of training samples')\n",
    "plt.ylabel('accuracy')\n",
    "plt.legend(loc='upper right')\n",
    "plt.ylim([0.48, 0.6])\n",
    "plt.tight_layout()\n",
    "# plt.savefig('./output/fig-learning-curve-%d.png' % d, dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkgYqSZIkqZL/AaTb3wnaia7mAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f0077d3e048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_sizes, train_scores, test_scores = learning_curve(estimator=clf, X=X_train_std, y=y_train, scoring=mse)\n",
    "\n",
    "train_mean = np.mean(train_scores, axis=1)\n",
    "train_std = np.std(train_scores, axis=1)\n",
    "test_mean = np.mean(test_scores, axis=1)\n",
    "test_std = np.std(test_scores, axis=1)\n",
    "\n",
    "plt.plot(train_sizes, train_mean,\n",
    "    color='blue', marker='o',\n",
    "    markersize=2,\n",
    "    label='Training error')\n",
    "plt.fill_between(train_sizes,\n",
    "    train_mean+train_std,\n",
    "    train_mean-train_std,\n",
    "    alpha=0.15, color='blue')\n",
    "\n",
    "plt.plot(train_sizes, test_mean,\n",
    "    color='green', linestyle='--',\n",
    "    marker='s', markersize=2,\n",
    "    label='Testing error')\n",
    "plt.fill_between(train_sizes,\n",
    "    test_mean+test_std,\n",
    "    test_mean-test_std,\n",
    "    alpha=0.15, color='green')\n",
    "\n",
    "plt.hlines(y=0.625, xmin=0, xmax=35000, color='red', linewidth=2, linestyle='--')\n",
    "plt.title('RandomForestClassifier')\n",
    "plt.grid()\n",
    "plt.xlabel('Number of training samples')\n",
    "plt.ylabel('accuracy')\n",
    "plt.legend(loc='upper right')\n",
    "plt.ylim([0.5, 0.7])\n",
    "plt.tight_layout()\n",
    "# plt.savefig('./output/fig-learning-curve-%d.png' % d, dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIq+R9gDjf95YOidwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f00775f9eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_sizes, train_scores, test_scores = learning_curve(estimator=tree, X=X_train_std, y=y_train, scoring=mse)\n",
    "\n",
    "train_mean = np.mean(train_scores, axis=1)\n",
    "train_std = np.std(train_scores, axis=1)\n",
    "test_mean = np.mean(test_scores, axis=1)\n",
    "test_std = np.std(test_scores, axis=1)\n",
    "\n",
    "plt.plot(train_sizes, train_mean,\n",
    "    color='blue', marker='o',\n",
    "    markersize=2,\n",
    "    label='Training error')\n",
    "plt.fill_between(train_sizes,\n",
    "    train_mean+train_std,\n",
    "    train_mean-train_std,\n",
    "    alpha=0.15, color='blue')\n",
    "\n",
    "plt.plot(train_sizes, test_mean,\n",
    "    color='green', linestyle='--',\n",
    "    marker='s', markersize=2,\n",
    "    label='Testing error')\n",
    "plt.fill_between(train_sizes,\n",
    "    test_mean+test_std,\n",
    "    test_mean-test_std,\n",
    "    alpha=0.15, color='green')\n",
    "\n",
    "plt.hlines(y=0.615, xmin=0, xmax=35000, color='red', linewidth=2, linestyle='--')\n",
    "plt.title('DecisionTreeClassifier')\n",
    "plt.grid()\n",
    "plt.xlabel('Number of training samples')\n",
    "plt.ylabel('accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.ylim([0.5, 0.7])\n",
    "plt.tight_layout()\n",
    "# plt.savefig('./output/fig-learning-curve-%d.png' % d, dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}