3.3 - RF - PNU All Data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SETUP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy as sp\n",
    "import matplotlib.pyplot as plt\n",
    "pd.options.display.max_rows = 400\n",
    "pd.options.display.max_columns = 400\n",
    "%matplotlib inline\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from epiml.loadepiml import LoadEpiml, save_search, load_search\n",
    "from epiml.epimlsklearn.epimlmetrics import pu_scorer, prior_squared_error_scorer_015, brier_score_labeled_loss_scorer, \\\n",
    "    f1_assumed_scorer, f1_labeled_scorer, report_metrics, f1_assumed_beta10_scorer\n",
    "from epiml.semisuperhelper import SemiSupervisedHelper\n",
    "from epiml.epimlsklearn.pnuwrapper import PNUWrapper\n",
    "from epiml.epimlsklearn.jsearchcv import JRandomSearchCV, extract_score_grid\n",
    "from epiml.epimlsklearn.nestedcross import NestedCV, rerun_nested_for_estimator, rerun_nested_for_scoring\n",
    "from epiml.epimlsklearn.frankenscorer import FrankenScorer, extract_scores_from_nested"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path = \"C:\\Data\\membership14_final_0103.txt\"\n",
    "lc = LoadEpiml(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.base import clone\n",
    "from sklearn.model_selection import validation_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(lc.X, lc.y, test_size=0.2, random_state=771, stratify=lc.y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Set up PNU Wrapper with Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pnu = PNUWrapper(base_estimator=RandomForestClassifier(), random_state=337)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Put in params that are decent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "params = {'base_estimator__class_weight': 'balanced',\n",
    " 'base_estimator__max_depth': 48,\n",
    " 'base_estimator__max_features': None,\n",
    " 'base_estimator__min_samples_leaf': 10,\n",
    " 'base_estimator__min_samples_split': 0.15,\n",
    " 'base_estimator__n_estimators': 56,\n",
    " 'num_unlabeled': 8303,\n",
    " 'pu_learning': True}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PNUWrapper(base_estimator=RandomForestClassifier(bootstrap=True, class_weight='balanced',\n",
       "            criterion='gini', max_depth=48, max_features=None,\n",
       "            max_leaf_nodes=None, min_impurity_split=1e-07,\n",
       "            min_samples_leaf=10, min_samples_split=0.15,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=56, n_jobs=1,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False),\n",
       "      num_unlabeled=8303, pu_learning=True, random_state=337,\n",
       "      threshold_set_pct=None)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pnu.set_params(**params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PNUWrapper(base_estimator=RandomForestClassifier(bootstrap=True, class_weight='balanced',\n",
       "            criterion='gini', max_depth=48, max_features=None,\n",
       "            max_leaf_nodes=None, min_impurity_split=1e-07,\n",
       "            min_samples_leaf=10, min_samples_split=0.15,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=56, n_jobs=-1,\n",
       "            oob_score=False, random_state=255, verbose=1, warm_start=False),\n",
       "      num_unlabeled=1.0, pu_learning=True, random_state=337,\n",
       "      threshold_set_pct=None)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pnu.set_params(num_unlabeled=1.0, base_estimator__n_jobs=-1, base_estimator__verbose=1, base_estimator__random_state=255)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train model on all data (num_unlabeled=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  4.5min\n",
      "[Parallel(n_jobs=-1)]: Done  56 out of  56 | elapsed:  7.9min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "PNUWrapper(base_estimator=RandomForestClassifier(bootstrap=True, class_weight='balanced',\n",
       "            criterion='gini', max_depth=48, max_features=None,\n",
       "            max_leaf_nodes=None, min_impurity_split=1e-07,\n",
       "            min_samples_leaf=10, min_samples_split=0.15,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=56, n_jobs=-1,\n",
       "            oob_score=False, random_state=255, verbose=1, warm_start=False),\n",
       "      num_unlabeled=1.0, pu_learning=True, random_state=337,\n",
       "      threshold_set_pct=None)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pnu.fit(X_train.values, y_train.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=8)]: Done  34 tasks      | elapsed:    0.0s\n",
      "[Parallel(n_jobs=8)]: Done  56 out of  56 | elapsed:    0.0s finished\n",
      "[Parallel(n_jobs=8)]: Done  34 tasks      | elapsed:    0.0s\n",
      "[Parallel(n_jobs=8)]: Done  56 out of  56 | elapsed:    0.1s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "({'SCORE': 0.85750315258511978,\n",
       "  'assumed_brier': 0.091438582251750583,\n",
       "  'assumed_brier_neg': 0.09135920365218797,\n",
       "  'assumed_f1': 0.095774647887323941,\n",
       "  'assumed_f1beta10': 0.70830411286663109,\n",
       "  'confusion_matrix_lab': array([[180,  35],\n",
       "         [ 78, 340]]),\n",
       "  'confusion_matrix_un': array([[81380,  6342],\n",
       "         [   78,   340]]),\n",
       "  'labeled_acc': 0.82148499210110582,\n",
       "  'labeled_avg_prec': 0.92164327233421772,\n",
       "  'labeled_brier': 0.13830509106280151,\n",
       "  'labeled_brier_neg': 0.19703509651483922,\n",
       "  'labeled_brier_pos': 0.10809707390445672,\n",
       "  'labeled_f1': 0.85750315258511978,\n",
       "  'labeled_prec': 0.90666666666666662,\n",
       "  'labeled_recall': 0.8133971291866029,\n",
       "  'labeled_roc_auc': 0.82530321575609211,\n",
       "  'pr_one_unlabeled': 0.072074234061275097,\n",
       "  'pu_mix_assumed_f1beta10': 79.557549326806821,\n",
       "  'pu_score': 8.7271380401437124},\n",
       " 0.85750315258511978)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "FrankenScorer()(pnu, X_test.values, y_test.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Validation Curve on num_unlabeled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  24 out of  24 | elapsed:  1.7min finished\n"
     ]
    }
   ],
   "source": [
    "param_range=[0,500,1000,2000,5000,7500,10000,15000]\n",
    "train_scores, test_scores = validation_curve(pnu, X_train.values, y_train.values, 'num_unlabeled',\n",
    "                 param_range=param_range, cv=3, scoring=f1_assumed_beta10_scorer, n_jobs=-1, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAisAAAGHCAYAAABxmBIgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmIhmyYoxxhhjIpolK8YYY4yJaP8Px7v2T96d5VMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xf0f13c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_scores_mean = np.mean(train_scores, axis=1)\n",
    "train_scores_std = np.std(train_scores, axis=1)\n",
    "test_scores_mean = np.mean(test_scores, axis=1)\n",
    "test_scores_std = np.std(test_scores, axis=1)\n",
    "\n",
    "plt.title(\"Validation Curve with assumed F1 Beta=10 for RF\")\n",
    "plt.xlabel(\"num_unlabeled\")\n",
    "plt.ylabel(\"F1 Beta=10\")\n",
    "plt.ylim(0.0, 0.8)\n",
    "lw = 2\n",
    "plt.plot(param_range, train_scores_mean, label=\"Training score\",\n",
    "             color=\"darkorange\", lw=lw)\n",
    "plt.fill_between(param_range, train_scores_mean - train_scores_std,\n",
    "                 train_scores_mean + train_scores_std, alpha=0.2,\n",
    "                 color=\"darkorange\", lw=lw)\n",
    "plt.plot(param_range, test_scores_mean, label=\"Cross-validation score\",\n",
    "             color=\"navy\", lw=lw)\n",
    "plt.fill_between(param_range, test_scores_mean - test_scores_std,\n",
    "                 test_scores_mean + test_scores_std, alpha=0.2,\n",
    "                 color=\"navy\", lw=lw)\n",
    "plt.legend(loc=\"best\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}