Active learnig by topics with dimension reduction.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from nltk.corpus import stopwords\n",
    "from gensim import corpora, models, similarities, utils, matutils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn import tree, metrics\n",
    "from sklearn.cross_validation import cross_val_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.naive_bayes import BernoulliNB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from libact.base.dataset import Dataset, import_libsvm_sparse\n",
    "from libact.models import LogisticRegression, SVM, Perceptron\n",
    "from libact.models import SklearnAdapter\n",
    "from libact.labelers import IdealLabeler\n",
    "from libact.query_strategies import RandomSampling, UncertaintySampling,HintSVM, QueryByCommittee\n",
    "from libact.query_strategies.multiclass import EER"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import copy\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>pid</th>\n",
       "      <th>title</th>\n",
       "      <th>abstract</th>\n",
       "      <th>topic_id</th>\n",
       "      <th>sr_title</th>\n",
       "      <th>rel</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3903849</td>\n",
       "      <td>[Nonvisualization of the gallbladder lumen by ...</td>\n",
       "      <td>Out of 12 000 sonographic examinations of the ...</td>\n",
       "      <td>CD011548</td>\n",
       "      <td>Ultrasound versus liver function tests for dia...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3903849</td>\n",
       "      <td>[Nonvisualization of the gallbladder lumen by ...</td>\n",
       "      <td>Out of 12 000 sonographic examinations of the ...</td>\n",
       "      <td>CD011549</td>\n",
       "      <td>Endoscopic ultrasound versus magnetic resonanc...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3903849</td>\n",
       "      <td>[Nonvisualization of the gallbladder lumen by ...</td>\n",
       "      <td>Out of 12 000 sonographic examinations of the ...</td>\n",
       "      <td>CD010339</td>\n",
       "      <td>Endoscopic retrograde cholangiopancreatography...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1559484</td>\n",
       "      <td>Cholangioscopy.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CD011548</td>\n",
       "      <td>Ultrasound versus liver function tests for dia...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1559484</td>\n",
       "      <td>Cholangioscopy.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CD011549</td>\n",
       "      <td>Endoscopic ultrasound versus magnetic resonanc...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       pid                                              title  \\\n",
       "0  3903849  [Nonvisualization of the gallbladder lumen by ...   \n",
       "1  3903849  [Nonvisualization of the gallbladder lumen by ...   \n",
       "2  3903849  [Nonvisualization of the gallbladder lumen by ...   \n",
       "3  1559484                                    Cholangioscopy.   \n",
       "4  1559484                                    Cholangioscopy.   \n",
       "\n",
       "                                            abstract  topic_id  \\\n",
       "0  Out of 12 000 sonographic examinations of the ...  CD011548   \n",
       "1  Out of 12 000 sonographic examinations of the ...  CD011549   \n",
       "2  Out of 12 000 sonographic examinations of the ...  CD010339   \n",
       "3                                                NaN  CD011548   \n",
       "4                                                NaN  CD011549   \n",
       "\n",
       "                                            sr_title  rel  \n",
       "0  Ultrasound versus liver function tests for dia...    0  \n",
       "1  Endoscopic ultrasound versus magnetic resonanc...    0  \n",
       "2  Endoscopic retrograde cholangiopancreatography...    0  \n",
       "3  Ultrasound versus liver function tests for dia...    0  \n",
       "4  Endoscopic ultrasound versus magnetic resonanc...    0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_all_topics = pd.read_csv('pubmed_data', usecols= ['pid', 'title', 'abstract', 'topic_id', 'sr_title', 'rel'])\n",
    "df_all_topics.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_reduced_data_by_topic(topic_id, df_all_topics, dictionary, lsi):\n",
    "    df_specific_topic = df_all_topics[df_all_topics['topic_id'] == topic_id]    \n",
    "    labels = []\n",
    "    corpus = []\n",
    "    stop_words = stopwords.words('english')\n",
    "    for index, row in df_specific_topic.iterrows():\n",
    "            if type(row['abstract']) == str: # revisamos si tenemos el abstract (no estan todos)\n",
    "                title = list(filter(lambda x : x  not in stop_words, utils.tokenize(row['title'], lowercase=True))) \n",
    "                abstract = list(filter(lambda x : x  not in stop_words, utils.tokenize(row['abstract'], lowercase=True)))\n",
    "                corpus.append(list(x[1] for x in lsi[dictionary.doc2bow(title+abstract)]))\n",
    "                labels.append(row['rel'])\n",
    "    \n",
    "    \n",
    "    return labels, corpus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "lsi = models.LsiModel.load('./lsi_model_600.model') \n",
    "dictionary = corpora.Dictionary.load('./dictionary_df_all_topics.dict')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_train_test(X, y, test_size, n_labeled):\n",
    "\n",
    "    X_train, X_test, y_train, y_test = \\\n",
    "        train_test_split(X, y, test_size=test_size)\n",
    "    trn_ds = Dataset(X_train, np.concatenate(\n",
    "        [y_train[:n_labeled], [None] * (len(y_train) - n_labeled)]))\n",
    "    fully_labeled_trn_ds = Dataset(X_train, y_train)\n",
    "\n",
    "    return trn_ds, X_test, y_test, y_train, fully_labeled_trn_ds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run(trn_ds, X_test, y_test, lbr, model, qs, quota):\n",
    "    Acurracy_out, F1_macro_out, presicion_out, recall_out, matthews_corrcoef  = [], [], [], [], []\n",
    "\n",
    "    for _ in range(quota):\n",
    "        ask_id = qs.make_query()\n",
    "        X, _ = zip(*trn_ds.data)\n",
    "        lb = lbr.label(X[ask_id])\n",
    "        trn_ds.update(ask_id, lb)\n",
    "\n",
    "        model.train(trn_ds)\n",
    "        y_pred = model.predict(X_test)\n",
    "        Acurracy_out = np.append(Acurracy_out, metrics.accuracy_score(y_test, y_pred))\n",
    "        F1_macro_out = np.append(F1_macro_out, metrics.f1_score(y_test, y_pred, average='macro'))\n",
    "        presicion_out = np.append(presicion_out, metrics.precision_score(y_test, y_pred, pos_label= 1, average='binary'))\n",
    "        recall_out = np.append(recall_out, metrics.recall_score(y_test, y_pred, pos_label=1, average='binary'))\n",
    "        matthews_corrcoef = np.append(matthews_corrcoef, metrics.matthews_corrcoef(y_test, y_pred))\n",
    "        \n",
    "\n",
    "    return  Acurracy_out, F1_macro_out, presicion_out, recall_out, matthews_corrcoef"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "topic_id = 'CD009944'\n",
    "labels, new_corpus = get_reduced_data_by_topic(topic_id, df_all_topics, dictionary, lsi)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_size = 0.33    # the percentage of samples in the dataset that will be\n",
    "# randomly selected and assigned to the test set\n",
    "n_labeled = 10   # number of samples that are initially labeled\n",
    "\n",
    "# Load dataset\n",
    "trn_ds, X_test, y_test, y_train, fully_labeled_trn_ds = \\\n",
    "     split_train_test(new_corpus, labels, test_size, n_labeled)\n",
    "trn_ds2 = copy.deepcopy(trn_ds)\n",
    "trn_ds3 = copy.deepcopy(trn_ds)\n",
    "lbr = IdealLabeler(fully_labeled_trn_ds)\n",
    "\n",
    "quota = len(y_train) - n_labeled    # number of samples to query\n",
    "\n",
    "\n",
    "model = LogisticRegression()\n",
    "qs = RandomSampling(trn_ds)\n",
    "Lg_R_Acurracy, Lg_R_F1_macro, Lg_R_presicion, Lg_R_recall, Lg_R_matthews_corrcoef = \\\n",
    "run(trn_ds, X_test, y_test, lbr, model, qs, quota)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = LogisticRegression()\n",
    "qs2 = UncertaintySampling(\n",
    "         trn_ds2, # Dataset object\n",
    "         model=LogisticRegression()\n",
    "     )\n",
    "Lg_U_Acurracy, Lg_U_F1_macro, Lg_U_presicion, Lg_U_recall, Lg_U_matthews_corrcoef = \\\n",
    "run(trn_ds2, X_test, y_test, lbr, model, qs2, quota)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f72efd55eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "query_num = np.arange(1, quota + 1)\n",
    "#plt.plot(query_num, Lg_U_Acurracy, 'm', label='uncertainty Acurracy')\n",
    "plt.plot(query_num, Lg_U_F1_macro, 'blue', label='uncertainty f1_macro')\n",
    "plt.plot(query_num, Lg_U_presicion, 'y', label='uncertainty Presicion ')\n",
    "plt.plot(query_num, Lg_U_matthews_corrcoef, 'red', label='uncertainty matthews correlation coef')\n",
    "plt.plot(query_num, Lg_R_matthews_corrcoef, 'orange', label='random matthews correlation coef')\n",
    "#plt.plot(query_num, Lg_R_Acurracy, 'g', label='random Acurracy')\n",
    "plt.plot(query_num, Lg_R_F1_macro, 'k', label='random f1_macro')\n",
    "plt.plot(query_num, Lg_R_presicion, 'brown', label='random presicion')\n",
    "\n",
    "plt.xlabel('Number of Queries')\n",
    "plt.ylabel('Score')\n",
    "plt.title('Logistic Regression Model Result (random vs uncertain)')\n",
    "plt.legend(loc='upper center', bbox_to_anchor=(0.5, -0.05),\n",
    "               fancybox=True, shadow=True, ncol=5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}