bagged_gbc.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# bagged_gbc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from __future__      import division\n",
    "from IPython.display import display\n",
    "from matplotlib      import pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import numpy  as np\n",
    "import pandas as pd\n",
    "import random, sys, os, re\n",
    "\n",
    "from sklearn.ensemble         import GradientBoostingClassifier\n",
    "\n",
    "from sklearn.ensemble         import BaggingClassifier\n",
    "\n",
    "from sklearn.cross_validation import StratifiedKFold\n",
    "from sklearn.grid_search      import RandomizedSearchCV, GridSearchCV\n",
    "from sklearn.cross_validation import cross_val_predict, permutation_test_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "SEED   = 97\n",
    "scale  = False \n",
    "minmax = False\n",
    "norm   = False\n",
    "nointercept = False\n",
    "engineering = False\n",
    "\n",
    "N_CLASSES = 2\n",
    "\n",
    "submission_filename = \"../submissions/submission_bagged_gbc.csv\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# Load the training data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from load_blood_data import load_blood_data\n",
    "\n",
    "y_train, X_train = load_blood_data(train=True, SEED   = SEED, \n",
    "                                               scale  = scale,\n",
    "                                               minmax = minmax,\n",
    "                                               norm   = norm,\n",
    "                                               nointercept = nointercept,\n",
    "                                               engineering = engineering)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "StatifiedCV = StratifiedKFold(y            = y_train, \n",
    "                              n_folds      = 10, \n",
    "                              shuffle      = True, \n",
    "                              random_state = SEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 60 ms, sys: 20 ms, total: 80 ms\n",
      "Wall time: 353 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "random.seed(SEED)\n",
    "\n",
    "\n",
    "# ------------------------------ base estimator ---------------------------------\n",
    "base = GradientBoostingClassifier(loss                     = 'exponential',\n",
    "                                  learning_rate            = 0.15, \n",
    "                                  n_estimators             = 175, \n",
    "                                  max_depth                = 1, \n",
    "                                  subsample                = 0.75, \n",
    "                                 \n",
    "                                  min_samples_split        = 2, \n",
    "                                  min_samples_leaf         = 1, \n",
    "                                  #min_weight_fraction_leaf = 0.0, \n",
    "                                  init                     = None, \n",
    "                                  random_state             = SEED, \n",
    "                                  max_features             = None, \n",
    "                                  verbose                  = 0, \n",
    "                                  max_leaf_nodes           = None, \n",
    "                                  warm_start               = False) \n",
    "                                  #presort                  = 'auto')\n",
    "\n",
    "# ---------------------------- BaggingClassifier ---------------------------------\n",
    "\n",
    "clf = BaggingClassifier(base_estimator     = base, \n",
    "                        n_estimators       = 10,\n",
    "                        \n",
    "                        max_samples        = 1.0, \n",
    "                        max_features       = 1.0, \n",
    "                        bootstrap          = True, \n",
    "                        bootstrap_features = False, \n",
    "                        \n",
    "                        oob_score          = False, \n",
    "                        warm_start         = False, \n",
    "                        n_jobs             = 1, \n",
    "                        random_state       = SEED, \n",
    "                        verbose            = 0)\n",
    "\n",
    "\n",
    "# param_grid = dict(max_samples        = [0.25, 0.50, 0.75, 1.0],\n",
    "#                   max_features       = [0.25, 0.50, 0.75, 1.0],\n",
    "#                   bootstrap          = [True, False],\n",
    "#                   bootstrap_features = [True, False])\n",
    "\n",
    "# grid_clf = GridSearchCV(estimator  = clf, \n",
    "#                         param_grid = param_grid,\n",
    "#                         n_jobs     = -1,  \n",
    "#                         cv         = StatifiedCV,\n",
    "#                         verbose    = 0\n",
    "#                        )\n",
    "\n",
    "# grid_clf.fit(X = X_train, y = y_train)\n",
    "\n",
    "# print(\"clf_params = {}\".format(grid_clf.best_params_))\n",
    "# print(\"score: {}\".format(round(grid_clf.best_score_, 4)))\n",
    "# print\n",
    "\n",
    "# clf = grid_clf.best_estimator_\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "clf_params = {'max_features': 0.75, 'max_samples': 0.5, 'bootstrap': False, 'bootstrap_features': False}\n",
    "clf_params['n_jobs'] = -1\n",
    "clf.set_params(**clf_params)\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# from sklearn_utilities import GridSearchHeatmap\n",
    "\n",
    "# GridSearchHeatmap(grid_clf, y_key='learning_rate', x_key='n_estimators')\n",
    "\n",
    "# from sklearn_utilities import plot_validation_curves\n",
    "\n",
    "# plot_validation_curves(grid_clf, param_grid, X_train, y_train, ylim = (0.0, 1.05))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/george/.local/lib/python2.7/site-packages/matplotlib/collections.py:590: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  if self._edgecolors == str('face'):\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmYAAAGJCAYAAAAg1v9AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2771116890>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 11.2 s, sys: 3.59 s, total: 14.7 s\n",
      "Wall time: 33.6 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "try:\n",
    "    from sklearn_utilities import plot_learning_curve\n",
    "except:\n",
    "    import imp, os\n",
    "    util = imp.load_source('sklearn_utilities', os.path.expanduser('~/Dropbox/Python/sklearn_utilities.py'))\n",
    "    from sklearn_utilities import plot_learning_curve\n",
    "\n",
    "plot_learning_curve(estimator   = clf, \n",
    "                    title       = None, \n",
    "                    X           = X_train, \n",
    "                    y           = y_train, \n",
    "                    ylim        = (0.0, 1.10), \n",
    "                    cv          = StratifiedKFold(y            = y_train, \n",
    "                                                  n_folds      = 10, \n",
    "                                                  shuffle      = True, \n",
    "                                                  random_state = SEED), \n",
    "                    train_sizes = np.linspace(.1, 1.0, 5),\n",
    "                    n_jobs      = 1)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training set predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.54 s, sys: 484 ms, total: 2.02 s\n",
      "Wall time: 5.65 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "train_preds = cross_val_predict(estimator    = clf, \n",
    "                                X            = X_train, \n",
    "                                y            = y_train, \n",
    "                                cv           = StatifiedCV, \n",
    "                                n_jobs       = 1, \n",
    "                                verbose      = 0, \n",
    "                                fit_params   = None, \n",
    "                                pre_dispatch = '2*n_jobs')\n",
    "\n",
    "y_true, y_pred   = y_train, train_preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[417  21]\n",
      " [100  38]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfQAAAHECAYAAADPrNgnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f27a4403450>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy 0.7899, mis-class rate 0.2101\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_true, y_pred, labels=None)\n",
    "print cm\n",
    "\n",
    "try:\n",
    "    from sklearn_utilities import plot_confusion_matrix\n",
    "except:\n",
    "    import imp, os\n",
    "    util = imp.load_source('sklearn_utilities', os.path.expanduser('~/Dropbox/Python/sklearn_utilities.py'))\n",
    "    from sklearn_utilities import plot_confusion_matrix\n",
    "\n",
    "plot_confusion_matrix(cm, ['Did not Donate','Donated'])\n",
    "\n",
    "accuracy = round(np.trace(cm)/float(np.sum(cm)),4)\n",
    "misclass = 1 - accuracy\n",
    "print(\"Accuracy {}, mis-class rate {}\".format(accuracy,misclass))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmUAAAGJCAYAAADL4URDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnACTMiIiIiIn8H+HFpEX0nAagQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f27a4403690>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.metrics import log_loss\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_true, y_pred, pos_label=None)\n",
    "\n",
    "\n",
    "plt.figure(figsize=(10,6))\n",
    "plt.plot([0, 1], [0, 1], 'k--')\n",
    "plt.plot(fpr, tpr)\n",
    "\n",
    "AUC = roc_auc_score(y_true, y_pred, average='macro')\n",
    "plt.text(x=0.6,y=0.4,s=\"AUC         {:.4f}\"\\\n",
    "         .format(AUC),\n",
    "        fontsize=16)\n",
    "\n",
    "plt.text(x=0.6,y=0.3,s=\"accuracy {:.2f}%\"\\\n",
    "         .format(accuracy*100),\n",
    "        fontsize=16)\n",
    "\n",
    "logloss = log_loss(y_true, y_pred)\n",
    "plt.text(x=0.6,y=0.2,s=\"LogLoss   {:.4f}\"\\\n",
    "         .format(logloss),\n",
    "        fontsize=16)\n",
    "\n",
    "f1 = f1_score(y_true, y_pred)\n",
    "plt.text(x=0.6,y=0.1,s=\"f1             {:.4f}\"\\\n",
    "         .format(f1),\n",
    "        fontsize=16)\n",
    "\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC curve')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIQAAAHuCAYAAADuhdetAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f27710bf690>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2min 29s, sys: 50.9 s, total: 3min 20s\n",
      "Wall time: 9min 25s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "score, permutation_scores, pvalue = permutation_test_score(estimator      = clf, \n",
    "                                                           X              = X_train.values, \n",
    "                                                           y              = y_train, \n",
    "                                                           cv             = StatifiedCV, \n",
    "                                                           labels         = None,\n",
    "                                                           random_state   = SEED,\n",
    "                                                           verbose        = 0,\n",
    "                                                           n_permutations = 100, \n",
    "                                                           scoring        = None,\n",
    "                                                           n_jobs         = 1) \n",
    "\n",
    "plt.figure(figsize=(20,8))\n",
    "plt.hist(permutation_scores, 20, label='Permutation scores')\n",
    "ylim = plt.ylim()\n",
    "plt.plot(2 * [score], ylim, '--g', linewidth=3,\n",
    "         label='Classification Score (pvalue {:.4f})'.format(pvalue))\n",
    "         \n",
    "plt.plot(2 * [1. / N_CLASSES], ylim, 'r', linewidth=7, label='Luck')\n",
    "\n",
    "plt.ylim(ylim)\n",
    "plt.legend(loc='center',fontsize=16)\n",
    "plt.xlabel('Score')\n",
    "plt.show()\n",
    "\n",
    "# find mean and stdev of the scores\n",
    "from scipy.stats import norm\n",
    "mu, std = norm.fit(permutation_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bagged_gbc                 ,        ,   0.7899 ,  7.2556 , 0.6137 , 0.3858 , 0.7573 , 0.0037\n"
     ]
    }
   ],
   "source": [
    "# format for scores.csv file\n",
    "import re\n",
    "algo = re.search(r\"submission_(.*?)\\.csv\", submission_filename).group(1)\n",
    "print(\"{: <26} ,        ,   {:.4f} ,  {:.4f} , {:.4f} , {:.4f} , {:.4f} , {:.4f}\"\\\n",
    "      .format(algo,accuracy,logloss,AUC,f1,mu,std))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# --------------------------------------------------------------------------------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Test Set Predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Re-fit with the full training set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "BaggingClassifier(base_estimator=GradientBoostingClassifier(init=None, learning_rate=0.15, loss='exponential',\n",
       "              max_depth=1, max_features=None, max_leaf_nodes=None,\n",
       "              min_samples_leaf=1, min_samples_split=2,\n",
       "              min_weight_fraction_leaf=0.0, n_estimators=175,\n",
       "              presort='auto', random_state=97, subsample=0.75, verbose=0,\n",
       "              warm_start=False),\n",
       "         bootstrap=False, bootstrap_features=False, max_features=0.75,\n",
       "         max_samples=0.5, n_estimators=10, n_jobs=-1, oob_score=False,\n",
       "         random_state=97, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.set_params(**clf_params)\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load the test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from load_blood_data import load_blood_data\n",
    "\n",
    "X_test, IDs = load_blood_data(train=False, SEED   = SEED, \n",
    "                                           scale  = scale,\n",
    "                                           minmax = minmax,\n",
    "                                           norm   = norm,\n",
    "                                           nointercept = nointercept,\n",
    "                                           engineering = engineering)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict the test set with the fitted model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 0 0 1 1 0 0 0 0]\n",
      "[[ 0.60294448  0.39705552]\n",
      " [ 0.80189223  0.19810777]\n",
      " [ 0.82778596  0.17221404]\n",
      " [ 0.69961581  0.30038419]\n",
      " [ 0.47036664  0.52963336]\n",
      " [ 0.09561834  0.90438166]\n",
      " [ 0.70877632  0.29122368]\n",
      " [ 0.9591084   0.0408916 ]\n",
      " [ 0.97481353  0.02518647]\n",
      " [ 0.98257413  0.01742587]]\n",
      "[0.39705552396159638, 0.19810777189126705, 0.17221404481928801, 0.30038419006148509, 0.52963336299652553, 0.90438165876083931, 0.29122367563814677, 0.040891601910195315, 0.025186473623049793, 0.01742586558028008]\n"
     ]
    }
   ],
   "source": [
    "y_pred = clf.predict(X_test)\n",
    "print(y_pred[:10])\n",
    "\n",
    "try:\n",
    "    y_pred_probs  = clf.predict_proba(X_test)\n",
    "    print(y_pred_probs[:10])\n",
    "    donate_probs  = [prob[1] for prob in y_pred_probs]\n",
    "except Exception,e:\n",
    "    print(e)\n",
    "    donate_probs = [0.65 if x>0 else 1-0.65 for x in y_pred]\n",
    "    \n",
    "print(donate_probs[:10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create the submission file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "assert len(IDs)==len(donate_probs)\n",
    "\n",
    "f = open(submission_filename, \"w\")\n",
    "\n",
    "f.write(\",Made Donation in March 2007\\n\")\n",
    "for ID, prob in zip(IDs, donate_probs):\n",
    "    f.write(\"{},{}\\n\".format(ID,prob))\n",
    "    \n",
    "f.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}