tfidf_with_ngrams_svm.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from nltk.stem import WordNetLemmatizer \n",
    "from sklearn.metrics import accuracy_score\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read the data\n",
    "\n",
    "Loading data into dataframe from 'citation_sentiment_corpus.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>Source_Paper_ID</th>\n",
       "      <th>Target_Paper_ID</th>\n",
       "      <th>Sentiment</th>\n",
       "      <th>Citation_Text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8731</th>\n",
       "      <td>W98-1118</td>\n",
       "      <td>W96-0213</td>\n",
       "      <td>p</td>\n",
       "      <td>He has achieved state-of-the art results by ap...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8732</th>\n",
       "      <td>W99-0606</td>\n",
       "      <td>W96-0213</td>\n",
       "      <td>o</td>\n",
       "      <td>B = (Brill and Wu, 1998); M = (Magerman, 1995)...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8733</th>\n",
       "      <td>W99-0607</td>\n",
       "      <td>W96-0213</td>\n",
       "      <td>o</td>\n",
       "      <td>The model we use is similar to that of (Ratnap...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8734</th>\n",
       "      <td>W99-0607</td>\n",
       "      <td>W96-0213</td>\n",
       "      <td>p</td>\n",
       "      <td>Our model exploits the same kind of tag-n-gram...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8735</th>\n",
       "      <td>W99-0608</td>\n",
       "      <td>W96-0213</td>\n",
       "      <td>o</td>\n",
       "      <td>In that table, TBL stands for Brill's transfor...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Source_Paper_ID Target_Paper_ID Sentiment  \\\n",
       "8731        W98-1118        W96-0213         p   \n",
       "8732        W99-0606        W96-0213         o   \n",
       "8733        W99-0607        W96-0213         o   \n",
       "8734        W99-0607        W96-0213         p   \n",
       "8735        W99-0608        W96-0213         o   \n",
       "\n",
       "                                          Citation_Text  \n",
       "8731  He has achieved state-of-the art results by ap...  \n",
       "8732  B = (Brill and Wu, 1998); M = (Magerman, 1995)...  \n",
       "8733  The model we use is similar to that of (Ratnap...  \n",
       "8734  Our model exploits the same kind of tag-n-gram...  \n",
       "8735  In that table, TBL stands for Brill's transfor...  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('citation_sentiment_corpus.txt', sep=\"\\t\", header=None)\n",
    "data.columns = [\"Source_Paper_ID\", \"Target_Paper_ID\", \"Sentiment\", \"Citation_Text\"]\n",
    "data.tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Removing duplicate data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original: (8736, 4)\n",
      "Drop Dupicates: (8734, 4)\n"
     ]
    }
   ],
   "source": [
    "print (\"Original:\", data.shape)\n",
    "data = data.drop_duplicates()\n",
    "data = data.reset_index(drop=True)\n",
    "print (\"Drop Dupicates:\", data.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adding lemmatized citation text to data frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "lemmatizer = WordNetLemmatizer() \n",
    "def lemmatize(cite_text):\n",
    "    result = []\n",
    "    for text in cite_text.values:\n",
    "        \n",
    "        #Lowercasing before negation\n",
    "        lower_case = text.lower()\n",
    "        #Lemmatization\n",
    "        lemm = lemmatizer.lemmatize(lower_case)\n",
    "        result.append(lemm)\n",
    "        \n",
    "    return result\n",
    "\n",
    "data['lemmatized_text']= lemmatize(data['Citation_Text'])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train and test split on citation text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "corpus = data['Citation_Text']\n",
    "num_training = 7000\n",
    "train_X = corpus[:num_training]\n",
    "test_X = corpus[num_training:]\n",
    "train_Y = data['Sentiment'][:num_training]\n",
    "test_Y=data['Sentiment'][num_training:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer = TfidfVectorizer(ngram_range=(1,3) ,min_df=1, use_idf=True, smooth_idf=True)\n",
    "\n",
    "# Vectorize the training data\n",
    "X_train = vectorizer.fit_transform(train_X)\n",
    "# Vectorize the test data\n",
    "X_test = vectorizer.transform(test_X)\n",
    "svm_C = 1000\n",
    "\n",
    "# Train the SVM, optimized by Stochastic Gradient Descent \n",
    "clf = SVC(probability=True, verbose=False, class_weight=None, kernel=\"rbf\")\n",
    "\n",
    "clf.fit(X_train, train_Y) \n",
    "\n",
    "# Make predictions\n",
    "pred = clf.predict(X_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print (accuracy_score(Y_test, pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import metrics\n",
    "\n",
    "matrix = metrics.confusion_matrix(Y_test, pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "\n",
    "def plot_confusion_matrix(cm,\n",
    "                          target_names,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=None,\n",
    "                          normalize=True):\n",
    "\n",
    "    import matplotlib.pyplot as plt\n",
    "    import numpy as np\n",
    "    import itertools\n",
    "    accuracy = np.trace(cm) / np.sum(cm).astype('float')\n",
    "    misclass = 1 - accuracy\n",
    "\n",
    "    if cmap is None:\n",
    "        cmap = plt.get_cmap('Blues')\n",
    "\n",
    "    plt.figure(figsize=(8, 6))\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "\n",
    "    if target_names is not None:\n",
    "        tick_marks = np.arange(len(target_names))\n",
    "        plt.xticks(tick_marks, target_names, rotation=45)\n",
    "        plt.yticks(tick_marks, target_names)\n",
    "\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "\n",
    "\n",
    "    thresh = cm.max() / 1.5 if normalize else cm.max() / 2\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        if normalize:\n",
    "            plt.text(j, i, \"{:0.4f}\".format(cm[i, j]),\n",
    "                     horizontalalignment=\"center\",\n",
    "                     color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "        else:\n",
    "            plt.text(j, i, \"{:,}\".format(cm[i, j]),\n",
    "                     horizontalalignment=\"center\",\n",
    "                     color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label\\naccuracy={:0.4f}; misclass={:0.4f}'.format(accuracy, misclass))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfwAAAHCCAYAAAAKInvOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeXxU1f3/8dcnmRAQZHeBgIKCIlCVxaVaLS5VWhDcQakVpbWtilv9Vqv+qrXu2rpUu1C1LrWCuAFqBfeqrQLiClpJBYWAIosgIEuSz++PeyZM9pDJZMjN++ljHsw999x7zk2cfOYs91xzd0RERCTecrJdAREREck8BXwREZFmQAFfRESkGVDAFxERaQYU8EVERJoBBXwREZFmQAFfZCuZWSszm2Zmq81schrnGWNmMxqybtliZoeY2X/reeyeZvaOmX1tZuc1dN1CGTuZ2b9CGb8zs8vM7O4a8i80syPDezOzv5nZKjObmYn6iTSGRLYrIJIpZnYqcBHQB/gaeAe41t1fS/PUJwI7AZ3cvbi+J3H3h4CH0qxLxpmZA73dvbC6PO7+KrBnPYv4JfCSu+9bz+PLMbOrgF7u/sOU5LOA5UBb3/rFR74DfA/o5u7rGqKOItmgFr7EkpldBNwGXEcUnHcB/giMbIDT7wp8nE6wjxMzS7fhsCswN8Nl7wrMq0ewTx67UMFemjx310uvWL2AdsBa4KQa8uQTfSFYEl63Aflh3xBgMfALYBmwFDgj7PsNsAnYHMoYB1wF/D3l3D0ABxJheyzwCVEvwwJgTEr6aynHHQTMAlaHfw9K2fcy8Fvg9XCeGUDnaq4tWf9fptT/WOAHwMfASuCylPz7A/8Bvgp57wRahH3/CteyLlzvqJTzXwJ8DjyYTAvH7B7KGBi2uwJfAkOqqOuLQAmwIZx/j/D7eyAc8ylwBZCT8jN7HbgVWAFcU+F8Qyv8ft4F7gvbm0LakVX8zk4LZa0ALgcWhnzjQt1KwrG/yfb/33rpVd9X1iugl14N/Qp/9IuTAbeaPFcDbwA7AjsA/wZ+G/YNCcdfDeSFQLke6BD2VwwWFbd7hCCZAFoDa4A9w74uQL/wfiwh4AMdgVUh8CSAU8J2p7D/ZeB/ISC2Cts3VHNtyfr/OtT/JyF4/gPYHugHfAP0DPkHAQeGcnsAHwIXpJzPibrIK57/RqIvTq1ICfghz0+AecB2wHTglhp+Fy8DP07ZfgCYEurag+hLyriUn1kxMD7Ut1UV5yv3+whp95Hy5SA1D9CXKJgfGq7n96GMIyv+nvTSqym/1KUvcdQJWO41d7mPAa5292Xu/iVRy/20lP2bw/7N7v4MUUCo7xh1KdDfzFq5+1J3r6r7ehgw390fdPdid38Y+Ag4JiXP39z9Y3f/BngEqGnMezPRfIXNwESgM3C7u38dyp8H7APg7m+5+xuh3IXAX4Dv1uGarnT3jaE+5bj7X4FC4E2iLzmX13I+AMwsFxgN/CrUdSHwO8r/bpa4+x9CfSuVXQ8nAk+5+7/cfSPw/4iuTyRWFPAljlYAnWsZ3+1K1IWb9GlIKztHhS8M64E2W1sRj8Z9RwE/A5aa2dNm1qcO9UnWqSBl+/OtqM8Kdy8J75NB8YuU/d8kjzezPczsKTP73MzWEM176FzDuQG+dPcNteT5K9Af+EMIpHXRmahXouLvJvXnsKiO56qrrqnnDL+zFQ1chkjWKeBLHP0H2Eg0bl2dJUSTsZJ2CWn1sY6o6zpp59Sd7j7d3b9H1NL9iCgQ1lafZJ2K6lmnrfEnonr1dve2wGWA1XJMjZPfzKwN0byIe4CrzKxjHeuynKh3ouLvJvXnUNvEu62dmLcU6J7cMLPtiHqJRGJFAV9ix91XE41f32Vmx5rZdmaWZ2bfN7ObQraHgSvMbAcz6xzy/72eRb4DHGpmu5hZO+BXyR3h/u+RZtaa6EvIWqruLn4G2MPMTjWzhJmNIhpbfqqeddoa2xPNM1gbeh9+XmH/F8BuW3nO24HZ7v5j4Gngz3U5KPRKPAJca2bbm9muRLdWbs3v5gugh5nV9e/bo8BwM/uOmbUgmruhv40SO/qfWmLJ3X9HFCiuIJqwtgg4F3gyZLkGmA28B7wPzAlp9SnrOWBSONdblA/SOaEeS4hmrn+XygEVd18BDCe6M2AF0Qz74e6+vD512koXA6cSzf7/K9G1pLoKuN/MvjKzk2s7mZmNJJo4mbzOi4CBZjamjvUZT9Rr8gnwGtFkw3vreCxAcjGkFWY2p7bMYU7DOaGcpUSTJRdvRXkiTYK51+e2VBEREWlK1MIXERFpBhTwRUREmgEFfBERkWZAAV9ERKQZUMAXERFpBvR43Azq3Lmz77prj2xXQ9JQqptYmryc2pYQkm3enDlvLXf3HTJdTm7bXd2L01ut2b/5crq7D22gKjUoBfwM2nXXHrz+5uxsV0PSsHFzSe2ZZJuWn5eb7SpImlrlWcVlpzPCi78hf89al5qo0YZ37qptWeqsUcAXEREBwKDOCzQ2PQr4IiIiED1BwuI7BqSALyIikhTjFn58r0xERETKqIUvIiKSpC59ERGRuNOkPRERkeYhxi38+H6VERERkTJq4YuIiEC4LS++7WAFfBERESAaw49vl74CvoiISFKMW/jxvTIREREpoxa+iIhIUoy79NXCFxERAcruw0/nVVsJZvea2TIz+6CKfb8wMzezzmHbzOwOMys0s/fMbGBK3tPNbH54nV6Xq1PAFxERgS0Pz0nnVbv7gKGVijbrDhwFfJaS/H2gd3idBfwp5O0IXAkcAOwPXGlmHWorWAFfREQkKcMtfHf/F7Cyil23Ar8EPCVtJPCAR94A2ptZF+Bo4Dl3X+nuq4DnqOJLREUawxcREWk4nc1sdsr2BHefUNMBZjYSKHL3d618L0EBsChle3FIqy69Rgr4IiIiQAOtpb/c3QfXuUSz7YDLiLrzM0pd+iIiIkk5lt5r6+0O9ATeNbOFQDdgjpntDBQB3VPydgtp1aXXfGn1qZ2IiEjsJJfWzeAYfkXu/r677+juPdy9B1H3/EB3/xyYCvwozNY/EFjt7kuB6cBRZtYhTNY7KqTVSAFfRESkkZjZw8B/gD3NbLGZjash+zPAJ0Ah8FfgbAB3Xwn8FpgVXleHtBppDF9ERCQpwwvvuPsptezvkfLegXOqyXcvcO/WlK2ALyIiAjTQpL1tlgK+iIhIkpbWFRERkaZMLXwREZEkdemLiIjEXN3Xw2+SFPBFRESSYtzCj++ViYiISBm18EVERJLUpS8iIhJ3ug9fRESkeYhxCz++X2VERESkjFr4IiIisOVpeTGlgC8iIgJoDF9ERKS50Bi+iIiINGVq4YuIiCSpS19ERKQZiHGXvgK+iIgIhIfnxLeFH98rExERkTJq4YuIiCSpS19ERCT+TAFfREQk3ox4B3yN4YuIiDQDauGLiIhAaOJnuxKZo4AvIiICgKlLX+JrxvRn2bvfnvTr04ubb7qh0v6NGzfyw1NH0a9PLw456AA+XbiwbN/NN15Pvz692Lvfnjw3Y3qt51y4YAGHHHQA/fr04oenjmLTpk0ZvbbmIpEDbfJzaJOfQ36i8h8rM2jdItrfukVOuQZMy4SVHdsyb8ue/ISxfX4ObVuW/xORG8pq2zKHhP56NJj6fg5XrFjB0UceRuf2bbjgvHPLHTNp4sMM3vdb7Ddgb0YMG8ry5csBeO/dd/nud77N4H2/xQnHHsOaNWsyfn1NiZml9dqW6SPbjJWUlHDBeecwZdo/efu9eUye+DAfzptXLs99995Dh/YdmPtRIePPv5DLL7sEgA/nzWPypInMeXcuU596lvPHn01JSUmN57z8sksYf/6FzP2okA7tO3Dfvfc0+jXHUcu8HNZtKmXtxlLyco2cCn9zWuXlsKnEWbuxlA3FpWWBPTcHcnOMtRujY3NzjNzwF6E45K+o1GH9plI2l3imL6vZSOdz2LJlS3591W+5/sZbyuUvLi7m/y46n2eff4lZb79H/2/tzZ//eCcAP//pj7nmuhuY/c77jBh5HLf+7ubGuVDJOgX8ZmzWzJnsvnsveu62Gy1atOCkUaN5atqUcnmemjaFMaedDsDxJ5zIyy++gLvz1LQpnDRqNPn5+fTo2ZPdd+/FrJkzqz2nu/PKSy9y/AknAjDmtNOZNvXJRr/muMnNiYKwh/i7ucTJyy0f8XMMikujDCWllNtfsUGSPE+JQ1Uh3T0qTxpOOp/D1q1bc/B3vkPLli3L5Xd33J1169bh7ny9Zg1dunQFoHD+x3znkEMBOPzI7/HkE481wlU2HWrhSywtWVJEt27dy7YLCrpRVFRUOU/3KE8ikaBtu3asWLGCoqLKxy5ZUlTtOVesWEG79u1JJKJpIwXdovySHiP6455U6pWDeEmpkxea/Ymc8AeNKPgXlzhtW0Zd9MUlrmCeBel8DquTl5fH7Xf+if0GfIvddunKhx/OY+yZ4wDYq28/pk2NvlA8/uhkFi9a1NCX1KQp4NeDmbmZ/S5l+2IzuyoD5VxWYfvfaZyri5k9VYd8z5tZh/qWI9KYNmx2ErnRWH0i1yh1x4la/jk5xpoNpazZUEoid0uXvjRtmzdv5q9/+RNvzHqbTz5bQv9v7c3NN14PwF/+ei8T/vxHDtp/EGvXfk2LFi2yXNttiDXAaxuWyY/3RuB4M+ucwTIAygV8dz8ojXNdBPy1DvkeBM5Oo5xtQteuBSxevOXbfVHRYgoKCirnCS2A4uJi1qxeTadOnSgoqHxs164F1Z6zU6dOrP7qK4qLi6P0xVF+SY9TfqGQHNvSLZ+aZ30Y49+wecvORK5RktKkLy5xcitOAJCMS+dzWJ1333kHgN123x0z48STTuaN/0RtoT379OGpf87g3zPf4uRRp9Bzt90b+pJkG5XJgF8MTAAurLjDzHYws8fMbFZ4HZyS/pyZzTWzu83s0+QXBjN70szeCvvOCmk3AK3M7B0zeyikrQ3/TjSzYSll3mdmJ5pZrpndHMp9z8x+mlK1E4BnQ/6xZva4mT1rZvPN7KaUfFOBUxrwZ5UVg/fbj8LC+SxcsIBNmzYxedJEhg0fUS7PsOEjeOjB+wF4/LFH+e5hh2NmDBs+gsmTJrJx40YWLlhAYeF89tt//2rPaWYcOuQwHn/sUQAeevB+hh8zstGvOW5KSiHXtnTj5+VapQl1qSE8P2FsKo72u0MiJcAncoxS9ek3unQ+h9XpWlDARx/O48svvwTgheefY88+ewGwbNkyAEpLS7nhumv4yVk/y8RlNUlGet3523qXfqbvw78LeK9CsAS4HbjV3V8zs12A6cBewJXAi+5+vZkNBcalHHOmu680s1bALDN7zN0vNbNz3X3fKsqeBJwMPG1mLYAjgJ+Hc6529/3MLB943cxmhGNWufvGlHPsCwwg6q34r5n9wd0XufsqM8s3s07uXm4gLXwZOQug+y67bN1Pq5ElEgluvf1Ojhl2NCUlJZw+9kz69uvH1Vf9moGDBjP8mBGMPXMcZ449jX59etGhQ0cefGgiAH379eOEk05mwN59SSQS3HbHXeTm5gJUeU6Aa6+7kdPGjOY3V17BPvsOKBtTlPR8s7mU1i2i7+6bwzh8fiJqvReXRhP7WuZF+0tKnW9CK39ziZfdZgfRxL7iMDG/ZcLIC7f4bd8yh03FzsZiJ9dgu/zo1r5EruFOlbP5pe7S+RwC7NmrB1+vWcOmTZuYNvVJnnpmBnv17ctlV1zJ9w4/lLxEHrvsuisT7rkPgEcmPsxf/nwXACOPPZ4fjT0jG5e9zdrWg3Y6zCv2/zXUic3WunsbM7sa2Ax8A7Rx96vMbBmwJCX7DsCewGvAce6+IJxjJbCHuy8P4//Hhfw9gKPd/Y1kOVWU2xL4GOgNDAVOdvcxZvYosDewPhzSDvgpsBb4tbsPDecZCxzs7j8J2/8ErnX318L268DP3P396n4GgwYN9tffnL31PzzZZmzcXJLtKkia8vNys10FSVOrPHvL3QdnupxEp9287Q+uSescq/4+plHqWh+NsdLebcAc4G8paTnAge6+ITVjdd+szGwIcCTwbXdfb2YvAy2rzBy4+4aQ72hgFJD8SmzAeHefnprfzAZUcc7U1n4J5X9eLYm+xIiISEzEuYWf8Tm57r4SeITy3fMzgPHJDTNLdsm/TtQNj5kdBSRnwrcj6m5fb2Z9gANTzrXZzPKqKX4ScAZwCGFsnmj44OfJY8xsDzNrTdQb0KMu12TR/xE7Awvrkl9ERCTbGusmnN8BqbP1zwMGh0lz84DkrJHfAEeZ2QfAScDnwNdEwTphZh8CNwBvpJxrAtE8gYeqKHcG8F3geXdPruN6NzAPmBPK+QuQcPd1wP/MrFcdrmcQ8Ia7F9chr4iINAWNcFuemd1rZstC/Emm3WxmH4WY+ISZtU/Z9yszKzSz/5rZ0SnpQ0NaoZldWqfLy9QYfn2ESXQl7l5sZt8G/lTNhLxMlX8cMMjdr6gl3+3AVHd/oaZ8GsNv+jSG3/RpDL/pa7Qx/M67efvh16V1jhX3n1JjXc3sUKI5Yw+4e/+QdhTRhPViM7sRwN0vMbO+wMPA/kBX4Hlgj3Cqj4HvAYuBWcAp7l5+TeaK15fWlTW8XYBHzCwH2AT8pDELd/cnzKz6m1u3+KC2YC8iIk1L8ra8THL3f5lZjwppM1I23wBODO9HAhPD3WMLzKyQKPgDFLr7JxDdhh7yNp2A7+7ziW6Dy2Yd7q5DnrosziMiIrK1ziSafwZQQPkh7MUhDWBRhfQDajvxNhXwRUREsqkBWvidzSx1LHeCu0+oY9mXEy1aV9WctLQp4IuIiCSl36O/vD7zDcLaL8OBI3zL5LoioHtKtm4hjRrSq6VHZYiIiABYdp6WF1aW/SUwwt3Xp+yaCowOK7v2JFpIbibRJL3eZtYzrCQ7OuStkVr4IiIijcTMHgaGEHX9LyZaUv5XQD7wXPjS8Ia7/8zd55rZI0ST8YqBc9y9JJznXKJ1ZXKBe919bm1lK+CLiIgEjTBLv6oHr91TQ/5rgWurSH8GeGZrylbAFxERCeK8tK4CvoiICI1zH342adKeiIhIM6AWvoiISFJ8G/gK+CIiIkDZbXlxpYAvIiISxDngawxfRESkGVALX0REJIhzC18BX0REJCm+8V4BX0REJCnOLXyN4YuIiDQDauGLiIhAWk+8awoU8EVERAIFfBERkWYgzgFfY/giIiLNgFr4IiIiSfFt4Cvgi4iIJMW5S18BX0REBGL/8ByN4YuIiDQDauGLiIgQDd/HuIGvgC8iIhLRwjsiIiLNQozjvcbwRUREmgO18EVERAJ16YuIiMSdxbtLXwFfRESEaJZ+Tk58I77G8EVERJoBtfBFREQCdemLiIg0A5q0JyIiEncxn7SnMXwREZFmQC18ERERkmvpx7eJr4AvIiICaC19ERGRZiLG8V4BX0REJCnOLXxN2hMREWkG1MIXEREB3ZYnIiLSHCRn6afzqrUMs3vNbJmZfZCS1tHMnjOz+eHfDiHdzOwOMys0s/fMbGDKMaeH/PPN7PS6XJ8CvoiISGCW3qsO7gOGVki7FHjB3XsDL4RtgO8DvcPrLOBPUR2tI3AlcACwP3Bl8ktCTRTwRUREGom7/wtYWSF5JHB/eH8/cGxK+gMeeQNob2ZdgKOB59x9pbuvAp6j8peISjSGLyIiEmRplv5O7r40vP8c2Cm8LwAWpeRbHNKqS6+RAr6IiEjQAPG+s5nNTtme4O4T6nqwu7uZedq1qIICvoiICIRZ+mlH/OXuPngrj/nCzLq4+9LQZb8spBcB3VPydQtpRcCQCukv11aIAr5IDXY+6PxsV0HStGrWndmugkhtpgKnAzeEf6ekpJ9rZhOJJuitDl8KpgPXpUzUOwr4VW2FKOCLiIiQvC0vw2WYPUzUOu9sZouJZtvfADxiZuOAT4GTQ/ZngB8AhcB64AwAd19pZr8FZoV8V7t7xYmAlSjgi4iIAI3x8Bx3P6WaXUdUkdeBc6o5z73AvVtTtgK+iIhIoJX2REREpElTC19ERCSI89PyFPBFREQg9g/PUcAXERFhy8Nz4kpj+CIiIs2AWvgiIiJBnFv4CvgiIiJBjOO9Ar6IiEhSnFv4GsMXERFpBtTCFxERAd2WJyIi0hxYI6yln00K+CIiIkGM473G8EVERJoDtfBFRESCnBg38RXwRUREghjHewV8ERERiIJ9nCftaQxfRESkGVALX0REJMiJbwNfAV9ERCQpzl36CvgiIiJBjOO9xvBFRESaA7XwRUREACNaXjeuFPBFREQCTdoTERGJO4v3w3M0hi8iItIMqIUvIiISxLiBr4AvIiIC0aQ9PTxHRESkGYhxvFfAFxERSYrzpL1qA76Zta3pQHdf0/DVERERkUyoqYU/F3AotwpBctuBXTJYLxERkUYVPR4327XInGoDvrt3b8yKiIiIZFucJ+3V6T58MxttZpeF993MbFBmqyUiItL4LM3XtqzWgG9mdwKHAaeFpPXAnzNZKREREWlYdZmlf5C7DzSztwHcfaWZtchwvURERBpds5yln2KzmeUQTdTDzDoBpRmtlYiISCOLFt7Jdi0ypy5j+HcBjwE7mNlvgNeAGzNaKxERkcYWHp6Tzqv2IuxCM5trZh+Y2cNm1tLMeprZm2ZWaGaTkr3oZpYftgvD/h7pXF6tAd/dHwCuAG4BVgInufvEdAoVERFpbsysADgPGOzu/YFcYDRRI/pWd+8FrALGhUPGAatC+q2k2diu69PycoHNwKatOEZERKRJSd6LX99XHSSAVmaWALYDlgKHA4+G/fcDx4b3I8M2Yf8RlsYkg7rM0r8ceBjoCnQD/mFmv6pvgSIiItuqTHbpu3sRUW/5Z0SBfjXwFvCVuxeHbIuBgvC+AFgUji0O+TvV99rqMmnvR8AAd18PYGbXAm8D19e3UBERkW1NA03a62xms1O2J7j7BAAz60DUau8JfAVMBoamXWId1SXgL62QLxHSREREpLzl7j64mn1HAgvc/UsAM3scOBhob2aJ0IrvBhSF/EVAd2BxGAJoB6yob8VqenjOrUS34q0E5prZ9LB9FDCrvgWKiIhsqzJ8H/5nwIFmth3wDXAEMBt4CTgRmAicDkwJ+aeG7f+E/S+6u9e38Jpa+B+Ef+cCT6ekv1HfwkRERLZlmQz37v6mmT0KzAGKiYbHJxDF2Ilmdk1Iuycccg/woJkVEjW+R6dTfk0Pz7mnun0iIiJxY5b5h+e4+5XAlRWSPwH2ryLvBuCkhiq71jF8M9sduBboC7RMqcgeDVUJERERyay63FN/H/A3op6O7wOPAJMyWCdpRDOmP8ve/fakX59e3HzTDZX2b9y4kR+eOop+fXpxyEEH8OnChWX7br7xevr16cXe/fbkuRnTaz3nwgULOOSgA+jXpxc/PHUUmzZtyui1NRffO2gv3n3i//HBlCu5+IzvVdq/S5cOPPPn8cyc9Cum//V8CnZsX7bvmvNGMnvyZcyefBknHjWwLH3I/nvw739cwhsTL+WFey9kt+6dAWiRl+DBG87ggylX8q8HLmaXLh0zf4HNQEN/DhctWsTRRx7GgL37MnCfftx5x+1l+VeuXMmwod+j/169GTb0e6xatSrj19eUNMJ9+FlTl4C/nbtPB3D3/7n7FUSBX5q4kpISLjjvHKZM+ydvvzePyRMf5sN588rlue/ee+jQvgNzPypk/PkXcvlllwDw4bx5TJ40kTnvzmXqU89y/vizKSkpqfGcl192CePPv5C5HxXSoX0H7rtXo0bpyskxbrv0ZEae+0cGnHANJw0dRJ/ddi6X5/oLj+Ohp2ey/6jruW7CP7l6/AgAhn6nH/vu1Z0DRt/AoafdwgU/OoLtW0edeHdcNpozLr+PA0ffwKR/zubSH0d3Do099tus+vob+o/8DX946CWuPX9k415wDGXic5hIJLjhpt/x9nvzeOW1N/jLn+8qO+ctN93AkMOP4IMP5zPk8CO4pYovGM1ZppfWzaa6BPyN4eE5/zOzn5nZMcD2Ga6XNIJZM2ey++696LnbbrRo0YKTRo3mqWlTyuV5atoUxpx2OgDHn3AiL7/4Au7OU9OmcNKo0eTn59OjZ092370Xs2bOrPac7s4rL73I8SecCMCY005n2tQnG/2a42a//j3436LlLCxawebiEiZPn8PwIXuXy9Nnty68MvO/ALwy62OGD/kWAHvttjOvzSmkpKSU9Rs28f78Io46aC8A3J22Ifi33b4VS79cDcDwIXvz0LQ3AXj8+bcZsv+ejXKdcZaJz2GXLl0YMDDqsdl+++3p02cvliwpKjvXD8O5fqjPYSXNvYV/IdCaaP3fg4GfAGdmslLSOJYsKaJbt+5l2wUF3SgqKqqcp3uUJ5FI0LZdO1asWEFRUeVjlywpqvacK1asoF379iQS0bSRgm7dyv4ASf113bEdi7/Y0iVb9MUqCnZoVy7P+x8XMfLwfQEYefg+tG3Tio7tWvPex1GAb9Uyj07tW/PdwXvQbecOAJx99T944g9nU/jsbzl12H7c8rfntpT3eVReSUkpa9Z+Q6f2rRvjUmMrE5/DVJ8uXMg777zNfvsfAMCyL76gS5cuAOy8884s++KLjFyXbHvq8vCcN939a3f/zN1Pc/cR7v56ugWbmZvZ71K2Lzazq+p5rvZmdnY9j11oZp3D+1Zm9oqZ5daQfwcze7Y+ZYlkw69ufYJDBvXiPw9fwiGDelH0xSpKSkp54Y2PePa1ebx03y+4//ozePO9BZSURE++Hj/mMI4b/0d6Df1/PDjlDW78xfFZvgqpj7Vr13LKySdw8+9uo23btpX2N4Vu6MZkGDmW3mtbVtPCO08QLbRTJXdP9y/ARuB4M7ve3Zenea72wNnAHyvuSFm9qC7OBB5395LqMrj7l2a21MwObogvPtnUtWsBixcvKtsuKlpMQUFB5TyLFtGtWzeKi4tZs3o1nTp1oqCg8rFdu0bHVnXOTp06sfqrryguLiaRSFC0eEt+qb8ly1bTbacOZdsFO3WgKHS/Jy39cjWjL74bgNatWnDsEfuyeu03ANx0z3Ruuiea6HXfdWOZ/9kyOndow7f2KGDWBzmuKbEAACAASURBVJ8C8OiMOUy56+wt5e3cgaJlX5Gbm0PbNq1Y8dW6jF9nnGXqc7h582ZOOfkERp0yhmOP2/LneseddmLp0qV06dKFpUuXssOOO2b4CpuQJtAtn46aWvh3AnfV8EpXMdGCAxdW3BFa0Y+Z2azwOjikX2VmF6fk+yA8H/gGYHcze8fMbjazIWb2qplNBeaFvE+a2VvhOcRnVVOnMYQVjixycyjjfTMblZLvyZC3SRu8334UFs5n4YIFbNq0icmTJjJs+IhyeYYNH8FDD0YPa3r8sUf57mGHY2YMGz6CyZMmsnHjRhYuWEBh4Xz223//as9pZhw65DAefyx6INRDD97P8GM04Stds+d+Sq9ddmDXrp3IS+Ry0tEDefrl98rl6dS+dVkr7v/OPJr7p0RrZ+XkGB3bRd3x/Xt3pX/vrjz/n49YtWY9bdu0otcuUSA4/MA+/HdB1O379CvvM+aYqGv4+CMH8MqsjxvlOuMsE59Dd+dnPxnHnn324vwLL6p0rr+Hc/1dn8NK4jxpr6aFd15ohPLvAt4zs5sqpN9O9Gzg18xsF2A6sFcN57kU6O/u+wKY2RBgYEhbEPKc6e4rzawVMMvMHnP3sjWJzawFsJu7LwxJxwP7AvsAncMx/3L3pURLIV5TVUXCl4mzALrvsktdfgZZk0gkuPX2Ozlm2NGUlJRw+tgz6duvH1df9WsGDhrM8GNGMPbMcZw59jT69elFhw4defChiQD07dePE046mQF79yWRSHDbHXeRmxuNhFR1ToBrr7uR08aM5jdXXsE++w5g7Jnjqq2b1E1JSSkX3vgI0/54Drk5xv1T3uDDTz7n//18GHPmfcbTr7zPoYN7c/X4EbjDa3MKueD6RwDIS+Ty/L0XAPD12g2cefn9ZV365/z2Hzx8y48p9VK+WvMNP73q7wDc9+S/ufeaH/HBlCtZtWYdp136t+xceIxk4nP4+muv8Y+HHqR//29xwKBo/sZvrrmOod//ARf/8lJ+eMrJ3P+3e9hll135+8OPZPPypRFZGsvyplew2Vp3b2NmVwObidYVbuPuV5nZMmBJSvYdgD2Bi4G17n5LOMcHwPCQ5yl37x/ShwBXuvthKeVdBRwXNnsAR7v7G2a2EBgMtCBap7hPyH8r8L673xu2HwQmu/tUM8sDPnf3Gh9TOGjQYH/9zdk1ZZFtXIf9zs12FSRNq2bdme0qSJpa5dlbNTyQpsHs2Ku/j7p5clrnuPP4vo1S1/qoy9PyMu02onWFU5sKOcCBYVnBMmZWTPlhiJZUr2xgMXwBOBL4truvN7OXqzj2m1rOl6plyC8iIjFhZPzhOVlVl9vyADCz/ExUwN1XEq3el9q/OwMYn1L2vuHtQqKuesxsINEzhQG+pua1AdoBq0Kw7wMcWEU9VgG5ZpYM+q8Co8ws18x2AA4FZoZ9e7Dl4UIiIhITOZbea1tWa8A3s/3N7H1gftjex8z+0MD1+B3ROHnSecBgM3vPzOYBPwvpjwEdzWwucC7wMUAYi389TLC7uYrzPwskzOxDogl+1T3xbwbwnfD+CeA94F3gReCX7v552HcY5Z8gKCIisk2rS5f+HUTj5E8CuPu7ZnZYzYfUzt3bpLz/AtguZXs5MKqKY74BjqrmfKdWSHo5Zd9GqlkO2N17pGzeRXTXwPPhmcP/F14VjQA0tVVEJGa29VZ6OuoS8HPc/dMK4xrV3qfelLn7HDN7ycxyq7sXP3Tv/z4MAYiISExEy+PGN+LXJeAvMrP9AQ8r0I0ndKXHUXJWfg37vyT0doiISLzEuYVfl0l7PwcuAnYBviCa8PbzTFZKREREGlatLXx3XwaMboS6iIiIZFWMe/RrD/hm9leqWFPf3atbnlZERKTJMdjmH4CTjrqM4T+f8r4l0Wp1i6rJKyIi0mTVeXGaJqguXfqTUrfDErOvZaxGIiIi0uDqs7RuT2Cnhq6IiIhItsW4R79OY/ir2DKGnwOsJHo6nYiISGyYWfMdw7doBYJ9gKKQVOrZeryeiIhIhsU43tcc8N3dzeyZ5GNnRURE4qy5L7zzjpkNyHhNREREJGOqbeGbWcLdi4EBwCwz+x/RM+aNqPE/sJHqKCIiknHN+T78mUTPnh/RSHURERHJqhjH+xoDvgG4+/8aqS4iIiLZY/Eew68p4O9gZhdVt9Pdf5+B+oiIiEgG1BTwc4E2hJa+iIhI3FmMQ15NAX+pu1/daDURERHJomjSXrZrkTm1juGLiIg0F3EO+DXdh39Eo9VCREREMqraFr67r2zMioiIiGSbxfi+vPo8LU9ERCR2mvMYvoiISPNh8V54py5r6YuIiEgDMLP2ZvaomX1kZh+a2bfNrKOZPWdm88O/HUJeM7M7zKzQzN4zs7SWtFfAFxERCXLM0nrVwe3As+7eh+jx8x8ClwIvuHtv4IWwDfB9oHd4nQX8Ka1rS+dgERGRuEiO4afzqvH8Zu2AQ4F7ANx9k7t/BYwE7g/Z7geODe9HAg945A2gvZl1qe/1KeCLiIgEZum9gM5mNjvldVbK6XsCXwJ/M7O3zexuM2sN7OTuS0Oez4GdwvsCYFHK8YtDWr1o0p6IiEjDWe7ug6vZlyB6Cu14d3/TzG5nS/c9ED173sw8ExVTC19ERAQAIyfNVy0WA4vd/c2w/SjRF4Avkl314d9lYX8R0D3l+G4hrV4U8EVERIjG8BugS79a7v45sMjM9gxJRwDzgKnA6SHtdGBKeD8V+FGYrX8gsDql63+rqUtfREQEoA4T7xrAeOAhM2sBfAKcQdT4fsTMxgGfAieHvM8APwAKgfUhb70p4IuIiDQSd38HqGqMv9Lza9zdgXMaqmwFfBERkaCO99I3SQr4IiIibBnDjysFfBERkSDOLXzN0hcREWkG1MIXEREJYtzAV8AXERGBsJZ+tiuRQQr4IiIiAAYW4yZ+nL/MiIiISKAWvoiISBDf9r0CvoiICBDG8GPcpa+ALyIiEsQ33GsMX0REpFlQC19ERCSIcY++Ar6IiEjEYn1bngK+iIgI8V94J87XJiIiIoFa+CIiIoG69EVERJqB+IZ7BXyRGs2dcXO2qyBpcvdsV0Gaipivpa+ALyIigibtiYiISAyohS8iIhKoS19ERKQZiG+4V8AXEREpE+MGvsbwRUREmgO18EVEREjO0o9vE18BX0REJIhzl74CvoiICACGxbiFrzF8ERGRZkAtfBERkUBd+iIiIjGnSXsiIiLNgcW7ha8xfBERkWZALXwREZEgzi18BXwREZEgzrflKeCLiIgQJu3FN95rDF9ERKQxmVmumb1tZk+F7Z5m9qaZFZrZJDNrEdLzw3Zh2N8jnXIV8EVERAJL8786Oh/4MGX7RuBWd+8FrALGhfRxwKqQfmvIV28K+CIiIoFZeq/az2/dgGHA3WHbgMOBR0OW+4Fjw/uRYZuw/4iQv14U8EVERIJGaOHfBvwSKA3bnYCv3L04bC8GCsL7AmARQNi/OuSvFwV8ERGRhtPZzGanvM5K7jCz4cAyd38rGxXTLH0REREabJb+cncfXM2+g4ERZvYDoCXQFrgdaG9midCK7wYUhfxFQHdgsZklgHbAivpWTC18ERERIP0O/Zq/Lbj7r9y9m7v3AEYDL7r7GOAl4MSQ7XRgSng/NWwT9r/o7l7fq1PAFxERgbK19DM5aa8alwAXmVkh0Rj9PSH9HqBTSL8IuDSdy1OXvoiISCNz95eBl8P7T4D9q8izATipocpUwBcREQlivNCeAr6IiAgkJ+3FN+Qr4IuIiATxDfeatCciItIsqIUvIiKSFOMmvgK+iIhIsBUPwGlyFPBFRESCGM/Z0xi+iIhIc6AWvoiISBDjBr4CvoiISJkYR3wFfBEREaJYH+dJexrDFxERaQbUwhcREYGyp+XFlQK+iIhIEON4r4AvIiJSJsYRXwFfREQEiKbsxTfia9KeiIhIM6AWvoiISKBJeyIiIjFnxHoIXwFfRESkTIwjvsbwRUREmgG18EVERALN0pfYmjH9Wfbutyf9+vTi5ptuqLR/48aN/PDUUfTr04tDDjqATxcuLNt3843X069PL/butyfPzZhe6zkXLljAIQcdQL8+vfjhqaPYtGlTRq+tuWiVl0O3Dvl075hPu1aVv8Mncowu7VpQ0CGfLu1akJvyqe/ZuSUFHfIp6JDPTm1bVDq2U+s8enRuWba9fctcuoX8Xdu3IC83vn8cG9OM6c+yT78+9N+rN7dU8zk87dTR9N+rN4cefGC5zyHAos8+Y4cO23Pb728pS+vTuyf7DdibAwYP4OAD9ytLP+3U0RwweAAHDB5An949OWDwgIxdV1Nklt5rW6aA34yVlJRwwXnnMGXaP3n7vXlMnvgwH86bVy7PfffeQ4f2HZj7USHjz7+Qyy+7BIAP581j8qSJzHl3LlOfepbzx59NSUlJjee8/LJLGH/+hcz9qJAO7Ttw3733NPo1x1Hn7fP4fPUmFq3cSJuWuZWCcMc2eXy9oYSiVRtZtb6Yjq3zyvY5ULRqI0WrNvLFmvJfwFokjJwKfyHWbixhccj/1fpiOrXJQ9JTUlLCheefy5PTnmHOu3OZPGli5c/h3+6hfYf2fPDhfMafdwFXXHZpuf2X/N8vOOro71c69z+fe5E3Z7/N62/MKkt78B8TeXP227w5+22OPe54Rh57XGYurImyNF/bMgX8ZmzWzJnsvnsveu62Gy1atOCkUaN5atqUcnmemjaFMaedDsDxJ5zIyy++gLvz1LQpnDRqNPn5+fTo2ZPdd+/FrJkzqz2nu/PKSy9y/AknAjDmtNOZNvXJRr/muMlP5LC5xCkudQDWbSihdYvccnla5BrfbC4FYMPm0kr7q9OpdR4r120ul+a+5b1t682ZJmL2rPKfmRNPHlXpc/j0tKn8MHwOjzvhRF5+KfocAkyd8iQ9evZgr759t6pcd+exRydz8qhTGuQ6ZNungN+MLVlSRLdu3cu2Cwq6UVRUVDlP9yhPIpGgbbt2rFixgqKiyscuWVJU7TlXrFhBu/btSSSiLueCblF+SU8iB4pLtkTh4lInt0ILf1NxKa3zo4/6di1yyMkxckIWAwra59O1fT7btdjy56Btq1zWbyqhpLRymW1b5tK9Yz6dWidYvnZz5QyyVZYUFVHQrVvZdvKzVDlP5c/h2rVr+f0tN3HZFVdWOq+ZccwPjuagAwZzz90TKu1//bVX2XHHnejVu3cDX1ETlm7zfhv/DtzoAd/MSszsHTP7wMwmm9l29TjH3WbWN7y/rMK+f6dRty5m9lQteVqY2b/MTBMepUlYsW4zrfJyKWifT6u83HJfED5buYGirzaybM0mOrXJI5Fj5OZAm/xcVn9TUuX51mwoYdHKjaxYV0yH7fQxyKZrf3sV48+7gDZt2lTa9/xLr/KfmW/x5LRnmPCnP/Laq/8qt/+RSQ9z8qjRjVXVJsPS/G9blo1P6zfuvi+AmT0E/Az4/dacwN1/nLJ5GXBdyr6D0qjbRcBfayl7k5m9AIwCHkqjrKzr2rWAxYsXlW0XFS2moKCgcp5Fi+jWrRvFxcWsWb2aTp06UVBQ+diuXaNjqzpnp06dWP3VVxQXF5NIJChavCW/1F9xKSRSWvSJHKMkJaADlJRSNj5vQOv8XMIIQFkLvrjU2bC5lPyEUYqRyM2he8f8smO6d8xn0cqN5c67bmMJO7TJ40vUyk9H14ICihYvLttO/SyVz1P5czhr5kyeePwxLr/sElZ/9RU5OTnkt2zJz88+t+yzvOOOO3LMyGOZPWsm3znkUACKi4uZ+uQTvPbG7Ma70CbA2PYn3qUj2136rwK9AMzsotDq/8DMLghprc3saTN7N6SPCukvm9lgM7sBaBV6DB4K+9aGfyea2bBkQWZ2n5mdaGa5Znazmc0ys/fM7Kcp9TkBeDbkH2tmU0JZ880stc/sSWBM5n4sjWPwfvtRWDifhQsWsGnTJiZPmsiw4SPK5Rk2fAQPPXg/AI8/9ijfPexwzIxhw0cwedJENm7cyMIFCygsnM9+++9f7TnNjEOHHMbjjz0KwEMP3s/wY0Y2+jXHzcbiUvJyjUToo2/dMpd1m8q3zHNS/oC13y7B1xuKK6XnGLRM5LCpxPlmUymfrdjAopUbWbRyIw5lwT71y8V2LaL5A5KeQYPLf2YefWRSpc/hD4Yfw9/D5/CJxx7lu0Oiz+HzL/2Lj+Yv4KP5Czhn/Pn83yW/4udnn8u6dev4+uuvAVi3bh0vPP8cffv1Lzvfiy88zx579qFbylCCxF/W+uNCl/j3gWfNbBBwBnAA0ZesN83sFWA3YIm7DwvHtEs9h7tfambnJnsMKpgEnAw8bWYtgCOAnwPjgNXuvp+Z5QOvm9mMcMwqd09txuwP9AfWA7PM7Gl3nw18AOxHE5dIJLj19js5ZtjRlJSUcPrYM+nbrx9XX/VrBg4azPBjRjD2zHGcOfY0+vXpRYcOHXnwoYkA9O3XjxNOOpkBe/clkUhw2x13kZsbTQar6pwA1153I6eNGc1vrryCffYdwNgzx2Xt2uNk+drN7NyuBWbw9YYSNpc4HbZLsLG4lPWbSmmVl0PH1nk40aS95Lh7Xm4OO2wfpRvw1TfFtQbwdi0TtGqRgwOlpc6yr3VrZboSiQS/v+0PjBg2lJLSEn50+hmVP4dnjGPc2B/Rf6/edOjQkQf+/nCN51z2xReMPul4IGrNnzz6FI46emjZ/kcfmcRJ6s6vUowb+Jh7435DN7MS4P2w+SrwC6JA3Mndfx3y/Bb4kqi1PYMoeD/l7q+G/S8DF7v7bDNb6+5tUs6/1t3bmFlL4GOgNzAUONndx5jZo8DeREEcoB3wU2At8Gt3HxrOMxY43N1/FLavBla6+21huwjo4+5fV7i+s4CzALrvssugj//3aQP81CRblqz6JttVkDR1ad+y9kyyTduuRc5b7j440+X032egT3721bTO0bdrm0apa31kdQw/qbrbe9z9YzMbCPwAuMbMXnD3q+tSiLtvCF8MjiYab5+YLA4Y7+7TU/Ob2QCg4l+Git+GUrfzgQ1VlDsBmAAwaNBg9XeKiDQh2/rEu3Rkeww/6VXgWDPbzsxaA8cBr5pZV2C9u/8duBkYWMWxm82sutU/JhENFRxCGJsHpgM/Tx5jZnuEMj8GelQ4/ntm1tHMWgHHAq+HYzoBy91ds5VERKRJ2CbuqXH3OWZ2HzAzJN3t7m+b2dHAzWZWCmwm6vqvaALwnpnNcfeKE+lmAA8CU9w9Odh4N1Fgn2NR18KXwLHuvtrM/mdmvdy9MOSdCTwGdAP+HsbvAQ4Dnk7zskVEZBsT51n6jR7wU8fbK6T/ngq354Vu9+lV5B2S8v4S4JKqzh9a4B0rHFtKdCtfufv3gzuBscAVYXuxux9bRb5TgUurSBcRkSYsxvF+22jhbyvc/YnQXV+tMOP/SXf/uJGqJSIijSXGEV8BvwJ3vzv8ex9wXxX7NwEPNG6tRERE0rOtTNoTERHJqmg5/MwtrWtm3c3sJTObZ2Zzzez8kN7RzJ4Li7w9Z2YdQrqZ2R1mVhgWiqtq4nqdKeCLiIgAbMVz76t71aIY+IW79wUOBM4Jz4W5FHjB3XsDL7Bljtj3idaS6U20vsuf0rk8BXwREZEgkw/Lc/el7j4nvP8a+BAoAEYC94ds9xPdBk5If8AjbwDtzaxLfa9NAV9ERKSRmVkPYADwJrCTuy8Nuz4HdgrvC4BFKYctDmn1okl7IiIiSenP0u9sZqmPIZwQVmDdUoRZG6I1Xi5w9zWpq826u5tZRlZpVcAXEREBqMPEuzpYXtNa+mGV18eAh9z98ZD8hZl1cfeloct+WUgvArqnHN4tpNWLuvRFRESCTE7aC6u73gN8GBabS5oKnB7enw5MSUn/UZitfyDRk16XUk9q4YuIiDSOg4HTgPfN7J2QdhlwA/CImY0DPiV6tDvAM0QPjyskesLrGekUroAvIiJC3Wbap8PdX6uhiCOqyO/AOQ1VvgK+iIhIkpbWFRERib8GmLS3zdKkPRERkWZALXwREZGgDsvjNlkK+CIiIkGM470CvoiICFD28Jy40hi+iIhIM6AWvoiISJn4NvEV8EVERAgL78Q33ivgi4iIJMU43ivgi4iIJMW5ha9JeyIiIs2AWvgiIiJBnJfWVcAXERFJim+8V8AXERFJinG81xi+iIhIc6AWvoiICNEM/TjP0lfAFxERCTRpT0REpDmIb7zXGL6IiEhzoBa+iIhIEOMGvgK+iIhIkibtiYiIxJ7FetKexvBFRESaAbXwRUREiMbv49ylrxa+iIhIM6AWvoiISKAWvoiIiDRpauGLiIgEcZ6lr4AvIiICoIfniIiIxJ8R75X2NIYvIiLSDKiFLyIikhTjJr4CvoiISKBJeyIiIs1AnCftaQxfRESkGVALX0REJIhxA18tfBERkTKW5qu205sNNbP/mlmhmV2agSuollr4IiIiQSYn7ZlZLnAX8D1gMTDLzKa6+7yMFZpCLXwREZHGsT9Q6O6fuPsmYCIwsrEKVws/g+bMeWt5qzz7NNv1yLDOwPJsV0LSot9h09Ycfn+7NkYhb895a/p2LaxzmqdpaWazU7YnuPuE8L4AWJSybzFwQJrl1ZkCfga5+w7ZrkOmmdlsdx+c7XpI/el32LTp99dw3H1otuuQSerSFxERaRxFQPeU7W4hrVEo4IuIiDSOWUBvM+tpZi2A0cDUxipcXfqSrgm1Z5FtnH6HTZt+f02Euxeb2bnAdCAXuNfd5zZW+ebujVWWiIiIZIm69EVERJoBBXwREZFmQAFfRKSZMYvzM+GkOgr4klX6wyOSFS1SN/Q5bB4U8CUrzGyn8EcmP2zr/8UmwswuNLOzzGyXbNdFtp6ZbQc8bmZnmNn3AdzdFfTjT7P0JSvM7CagFZAH/Mnd381ylaSOzGwAcCawDmjl7udnuUqylcxsf2AvYCiw3t3HhXRzBYXYUsCXrDGzAuD7wKXABcCz7l6c3VpJXYRWYlfgSqCHux+S5SpJHSRb8aFFn0PUw/Yc8Jm7n5rMo6AfTwr4knVmdgJwPnCbuz9uZjnuXprteskWZpbr7iUV34ftR4DW7j4saxWUWiUDuZl1Bta5+zcp+14DPnH3H2WvhpJpGjeVRpNsXZhZHzM7MJnu7o8BtwI3mdkgBfttS/gCVmJmOWEo5k9mdnLooQEYBRSZ2U+zWE2pRQj2xwCTgWfN7Mdmtk/YPQRob2anZq2CknEK+NJowh+co4Engb+a2Z1mtntoeTwBXA9cYmbtsltTSeXupeHL2uNEy3G/BpwKjDKzzqH7dwqwYxarKbUws4HAL4h60+4AegAjzGynMJR2D9A+ezWUTFPAl4xKnX1vZv2Bc4nG7QcAbYGfAruHLJOAT4gm80mWmdlpKZt7ANu5+0Xu/gDwW6IJXweH/bOAg8KEPtkGmFk3Mzs9vN8ZOAdIuPt7oVftaeAQtjxr/hNgv5BXYkgBXzLGzNoTteS3N7M2wPHA3kDH0KK4iKhVeL6Z7ebua4ElQP+sVVpSDTezIeF9EbDczE40s+3c/S3gUeCAMKa/jOjLXGGW6iqVJYCB4Uv3BuB5ADM7D8Dd/wPMBwaF7feJJmGuyUptJeMU8CVj3P0r4HOiCV1ribp9HwPOMLN+7r4cuBjoRGjVu/sdwL+zVGUp7xWgV3i/CZgDDASON7OWwNHAqpQJfJ+4+9eNX02pxiaiZ6/vGz6L04A/AHub2V1mdghwJPBB8gB3/8zd12eltpJxmqUvGRNaFncQDd+PD2mDgOFAB+Aed3/fzPLcfbNm52efmV0LTHH3mWHM9xHgYnd/0sy2J7r/fhBRz8xCd/9ZFqsrtTCzs4CzgZPcfX7odTuS6FbY5cCN7v5SxTsvJJ4U8CWjzKwVMAN42t1vCGkDgROBdsAlRAt/KNBvA8zs18ApwBh3nxNmdd9CFPSnmVkCcKC7+/9v77zDrSquPvz+ABUrajSKmAR7V2IPiX7EXhLEFtFosNdYE2OJyad+tmiMosbHQhRNFAsiErvGGrtSNbGXqLGgYANFwd/3x5qjmysXLhfkHLnrfZ77cM7s2TOz9+bsNWvNmrX8cjknJ2oNQmXr3XzFqoako4D1gCNtv1AmblsR/hdP2b6wjkNOZiEd6j2AZPai8sLpDMxp+xVJRwDHSPqV7TOLIPkc+Lj2UkrqS01o2z5R0jjgSkk/L0J+buD88u+gogm+XM5TCvvGofz2egK9JE0CLiO24XUArpG0h+2RkgYSUS7XkrSQ7bF1HHYyi0gNP5npSOpFRM77AHgK6Ad0Bk4BhhNa/STbn6Z2WH9q5tzqs5D0a2BvYJcyQdsE+A3wCDDG9ll1HHLSDJK6A+cBPYErgdeAPWxPkHQAYVm7C3gAuI/wr0m/izZCCvxkpiJpNeDPhMnwUGAboIft8cWUeCHwCjAe+GM12lcy66kKe2LXBEA/2+9JOpTYNvkL24+XQDvzElu5Btp+v07DTppB0p6ER/5bwMlAb9svS+pk+31JqxAOsl2BIbY/rd9ok1lNCvxkpiJpZcIpbzSwD7Cr7RcldbM9XFJ7IrjHksALadKvP0XY30xofGsQnt1H2H64CP3jge/X1uyTxkPSCuXj0oR1bSHgZ0XY9yYmaYdmroq2TW7LS2YKklaWtCOxFWgDwjP4F0XYb0msAS9he5Ltd22PSGFfP8rEq8YhROCcU4ktkh8CJ0taz3ZfYKcU9o2JgjmJCHo/Jcz1nxHbXztKWg84Frg5hX2SGn4yU5C0D7FW2F3SYcC6xMtnPPBb4CjbN9ZzjElQMePPAfzS9lklMFJfYLjtcyXdS2RS29X28+W89LdoECR1qApwSasDVwA7Eib7PkQEvbmAC2wPqTnU1mXASUOQXvpJq6h443ewPdH2xZJ6SDrY9tmS9iZeOAsDh9u+PV84jUER9gLOBOYsZR8VPycyywAAErFJREFUrb/mwPU0sc/++cp5KezrjKSlgVdL3IrVgUWAZ4rn/V+BDW1fJOlFYrLdyfY7+dtLIAV+Mp1IWh5Yw/a1JYjOjyU9b3swcCkRfQ3b/Ur9OWx/VsryhdMAFGG/G/AzoFsp60DES9hF0sHAE7ZPrdXPZ9cw7AbcBjwM/ICIfLiipF8BbwO7ShpQcah8B/K3lwS5hp9ML+2At4vH/WuEhniQpHOBicCWmjzpSq4bNgCS1pX0nfJ1fyKG+lPAmRXz8O3EWvCZtvct56WwbyBsnwC8LGk4cInt/YABxDr9t4EfE7kpVMdhJg1KCvxkurD9NLGH91Wgl+1TiD2/7YloXgsCfcqacGoWjUMH4BZJjwFzOxKnHAy8D5xdhP47tp+xfSV8sWafz6/OKJJPLVU+r2n7TWLJ5VFJHW1fAPweeIjId/BwPrdkSqTTXjJNJM0DbGr7huL1+ykg4FbgZNt9y9auxQkz8XO2b6rfiJMaqsRIL454KwBb235CkQBnGSLL3UJEON2Mp95glHX7a4l0tptRHCklXUpMsr9ve0KpW/OtSctM8hVS4CctQlJ/YG0iqMc+tocpYuLfCRxn+/wm9fOF0yAU8+73CJPvUoT598iaIyURU72L7avrOMykCZK+CyxR4iEcDZwE/Nb2Hyp1LiGS4SxXounl7y5pljTpJ1OlshZ4KuFxP9H2MADbQ4mXTd8SoOUL8qXTUOxCTMxchHpf4pltBJxFJMK5GiZ73kn9WQv4TNK8wAgioM6xilj5ANjek0g7vX75nr+7pFlSw0+apWIebAfMR5h9LwE+s71Fpd5yQFfbd9RpqEmFKWl5itj4OwIHlTC5uxJr+MOcKW4bFkU622uBs23fJGkb4HJga+A9YhnmmFI3tftkqqTAT6ZIRdhvRmgPb9q+qBy7CxhHmBhPB7a1PSZfOI2FpIOAt2wPLN9/BewL9Clm4kVsv1OO5bNrECq/vc6235C0P7Ad8Cfbt0raloih8B5wYtkSmyTTJPfhJ1OkvHC2IF4svwQGSFoD+J3tjSQNAE4gtnCNqZ1TvxEnTRz0FiIyFC4j6RPbN9o+U9L6wHWSfugv89mnsG8QKsK+J7CPpMNsX6BIdXuUJGxfL+lxoIPtl/L5JS0lBX7yFYoJf35iv3ZvYDHgDcLx65wSTW9nSQs6sqrlC6fOaPKsd2sB42wfJ+lIYINy/AbgWWCEK7Hx89k1DkXYb0BMpvvYfqEEr7pY0tvAieX3dkv1nLoNOPlGkQI/+YKK4O7oSKW5F+GodyLhoT838CbwqqQTbb8H+cKpN+W51YT97cCjwM8lnUNEP9wN6C3pFOBB2/tUzstn13isCNwLTJJ0IBHMahywJ7AoMKaeg0u+uaSXfgJMZkpcD7hf0mq23yUmhZ8SDntdiIQ4g5x57BsCSXNWHCvPBG63fSyRnvjjskZ/PnAosRUvhX2DUdsZUeIiADwCzAsMLN8vAP4DLGO7n+1HZv0ok9mB1PATYLI1+x2IqHm3Sdrc9ihJjxKZuJYFDrT9WD3HmgSSdga6SjrH9jhJTwGvS7oTGGj7fEkrAovZvpfIeZ9Z7xqMym9vG0kTgUHA74D3bX8s6ftE6tsB9Rxn8s0nNfwEAEXozr5EfO5uhFZxg6SVgOOINcWdbd9Wx2EmhaIVzkX4V+wpaS5iAj8IuMr2aaXqqUD36rkp7BuDima/PmGFGQxMIrbc9QTmkPQ/wNWEdWZYvcaazB7ktrwEAEkLAOcBv3HE6kbS5UAP4Ie2X63j8JIKkhYu2yDbEYJhU2Ck7Qsl9QXWAa4hgiK9a7tPHYebNEHS4sA8xJLLG4pkU2vZPqwc34mwtB1A7LTomFa1ZGaQGn4bpaJddJLUyfYHwALEft8aVxBrwTeoJMNJ6ouky4gkOLsQQmIwkS51WUn72z6U0BY/BG6tCfsyOUjqTFliGQJcCJws6SfAv4BVFemmKVEPFyDC5Y5KYZ/MLHINv41S1g1/ChwBjJX0MHA0sd9+SeBjQvjvAexHOBF9VK/xJl9wFZFEpSfwjqTjCVPwBGA5SbsDV9r+tHZCrtk3BpJWJibRRxDpiXsRa/NnEhkoexbt/0VgCWBsnYaazKakSb8NUfXMLuuGZxHhVncF9rK9nKTlCVPwd4mX0yLE2v7GtkfXZ+RJFUk9gMuAVYCNiAx4+wImntvmxUkvaSAk/Qi4z3a78n1F4je4A7A08CMi7sWHhC/NoHqNNZk9SYHfRpC0KKFRDLD9kaQNidz1cxEaxy4lalfXSgS27kTc7l62n6zT0JMpIGlr4Aygewl+tCIRM2Ez28fXdXBJsxRv/PNtL11C5B5EhKb+sByfF5grQ1UnXwcp8NsIknoBPwGGA/0Jx67zgHeBnkVobEpE19u/lHcmwne+UpdBJ1NF0lbAOcC6tfDGlWNpxm9QynO7Bnga2ND2+JpPTQr45OskHXlmcyS1Lx//DtxCmH93s303sYVrYaBz8Qw+mzAljrb9ue3XU9g3LrZvJjLevSppnibHUtg3KOW59QQWtz2+FLdPYZ983aSGPxsjaQVgbyLc6n22J0jaEtgS+FdJynE8ockvSAj729KU+M1C0uq2R9Z7HMn0Ucz7lwMr2E4HveRrJwX+bEwJ2nE34RF8DeEYdAaxb3tO4L9A/+Kx39H2J3UbbDLDpBn/m0cx74+3fU+9x5LM/qTAn80pnsE3AusB2xMx8bcFXiNC5R4PXAJpBk6SepFWtWRWkPvwZ3Ns/7PEXB9IeHR/KOlGYDViK9dLKeiTpL6ksE9mBanhtxGK6fBcYJ2aR3clQ15qF0mSJLM5qeG3EWzfLOlz4GlJK9geWxPyKeyTJElmf1LDb2OUgC3j0kkoSZKkbZECv42SZvwkSZK2RQr8JEmSJGkDZKS9JEmSJGkDpMBPkiRJkjZACvwkSZIkaQOkwE+SWYikSZKGS3pS0rVNk95MZ1s9ShAlJPWUdPRU6i4o6cBW9HG8pF+3tLxJnf6SdpiOvrpKyjTMSfI1kQI/SWYtH9vuZntV4FMiFfEXKJju36XtIbZPm0qVBYHpFvhJksw+pMBPkvpxP7Bs0WyfkXQ58CTwHUmbSXpI0tBiCZgPIsOapKclDQW2qzUkaXdJ55XPi0m6XtKI8tcdOA1YplgXzij1jpT0mKSRkk6otPVbSc9K+ieRTnmqSNqntDNC0nVNrBabSHq8tPeTUr+9pDMqfe83ozcySZJpkwI/SeqApA5EmuJRpWg54HzbqwDjgOOATWyvCTwOHCGpI3Ax8FNgLWDxZpo/B7jX9hrAmsBTwNHAC8W6cKSkzUqf6wLdgLUkbShpLaB3KdsKWKcFlzPI9jqlv38De1WOdS19bA1cUK5hL+B92+uU9veRtFQL+kmSZAbI0LpJMmuZW9Lw8vl+4C/AEsArth8u5esDKwMPSIJIZfwQsCKR7Og5AEl/IxIgNWUj4BcAticB70taqEmdzcrfsPJ9PmICMD9wve3xpY8hLbimVSWdRCwbzAfcVjl2TUnO9JykF8s1bAasXlnf71T6frYFfSVJ0kpS4CfJrOVj292qBUWoj6sWAXfY3rlJvcnOm0EEnGr7wiZ9HNaKtvoDvWyPkLQ70KNyrGlkL5e+D7ZdnRggqWsr+k6SpIWkST9JGo+HgR9KWhZA0rySlgeeBrpKWqbU27mZ8/8BHFDObS+pE/Ahob3XuA3Ys+Ib0EXSt4H7gF6S5pY0P7F8MC3mB96QNAfw8ybHdpTUrox5aeCZ0vcBpT6Slpc0bwv6SZJkBkgNP0kaDNuji6Y8QNJcpfg4289K2he4SdJ4Yklg/ik0cShwkaS9gEnAAbYfkvRA2fZ2S1nHXwl4qFgYPgJ2tT1U0tXACOBt4LEWDPl3wCPA6PJvdUz/AR4FFgD2t/2JpH7E2v5QReejgV4tuztJkrSWjKWfJEmSJG2ANOknSZIkSRsgBX6SJEmStAFS4CfJLETSXJKulvS8pEea80yXdLikp0oI3gFl/3otXO1LJYDO8JrnvqSFSrCdkZIelbRqKe9Yvo8o7Z0wpf5aeS39JK08nefM8vC5ko4p9/sZSZs3U+eXpY4lLTKF4+tImljZSoikPpKeK399Stn8lWczXNI7ks7++q4uSVpOOu0lbR5JHWxPnEXd7QWMtb2spN7AH4CdmoynC3AIsLLtjyVdQwTD6V+qHGl7YJN2jwWG295W0orAn4GNgQnARrY/Kl7x/5R0S2XPf6uxvfeMtvF1UyYkvYFViHgHd0pavsQnqPIAcCNwzxTaaE88p9srZQsD/wusTWw1fELSENtjiaBFtXpPAINm5jUlSWtJDT9pWCQNlvRE0Uz3rZRvoQg5O0LSP0rZfJIulTSqaLnbl/KPKuftIKl/+dxf0gWSHgFOl7SuIpTtMEkPSlqh1Gsv6Y9F0x4p6WBJG0kaXGl3U0nXt/CytgEuK58HAhsXT/WmdCCC9HQA5gH+O412VwbuArBd2763mIPaPZij/LmM+0RJPZs2pEiMc5mk+yW9Imk7SaeXe3trZTvdPZLWLveof7lHoyQdXo4vK+nO8pyG6svthLV+upY+hpa/7qW8s6T79GWSoQ2a66MFbANcZXuC7ZeA54nIf5Nhe5jtl5tp42DgOmLXQo3NiVgJY4qQvwPYosn1LQ98m9hNkSR1JzX8pJHZ0/YYSXMDj0m6jpikXgxsaPulomlBbA173/ZqECbuFrS/JNDd9iRJCwAb2J4oaRPgFGB7IpJdV6BbObYwMBY4X9KitkcDewCXlH6vZsrx5/9k+3KgC/AqQGnvfeBbwDu1irZfl/RHYkvbx8Dttm+vtHWypN8T++2Ptj2B2Ea3HXC/pHWB75Xre6toqE8AywJ/tv1I6ef3U7k3ywA/JiYSDwHb2/5NmdhsDQyu1O0GdCkJgZC0YCm/AjjN9vWKJYl2hACs8TawadmqtxwwgNCYdwFus31yGfs8zfUh6Ui+uvcf4D7bh5T7XbVmvFbKWkSxtmxb7kU1zPAXz3Eq7fYGrnZuhUoahBT4SSNziKRty+fvEOFXFyVe5i8B2B5Tjm9CvGAp5WNb0P61FdNuJ+CyInhMaMK1di+omfxr/Un6K7CrpEuBH/BlKNvJzPOtoUxWtgGWAt4DrpW0q+2/AccAbxLhdi8CjgJOJJLj9FWE7R1FhMydVMY0CehWhOT1kla1Pa119FtsfyZpFNAeuLWUjyImQFVeBJaWdC5wE3C7ImhPF9vXlzF8Uq6tet4cwHkKP4RJwPKl/DHgkmJJGGx7uCIs72R9lHbPAM6YxrXMCGcDR9n+fMqGmKnSG9ht5g8pSVpHCvykIZHUgxC2P7A9XtI9QMdWNFXVrpqeXw1n+3/A3WUNvCtTWMttwqXA34FPiInDxDLuaWn4rxOTl9eKub4T8G6TupsQMfNHlzYHAd2Bv9l+o9SZUCYbvwaw/QFhaaAsEbxECOIvsP2epLsJ0/O0BP6Ecs7nkj6raKmf0+S9YXuspDUIM/f+wM+I4D/T4nDgLWANQvv/pLR3n6QNCUtCf0l/sn35FPrYswUafu1+11iylLWUtYGrirBfBNhK0sTSRo8m7d5T+1LG2sH2E9PRV5J8reQaftKodCKc28YrnNDWL+UPAxuqZFermPTvAA6qnVwx6b8laSVFjvmataC5/mqCYPdK+R3AfkU4f9Gf7f8S6+rHEcKfUr5TyUjX9O/yUmUI0Kd83gG4awom3/8A60uapwjvjYksdEjqXP4VEZ3uyfJ9QUlzlvP3JgTeB5IWrZi/5wY2JUL0IunUigWl1Si82tvZvq7cjzVtf0hManqVOnNp8rS5EPf8jZJcZzfCkoCk7wFv2b4Y6AesOaU+IDT8Zu73IaWPIUDv0v9ShJXo0ZZem+2lbHe13ZXwuTjQ9mAiPPBmit0RCxEJgaq5AXYmliiSpGFIgZ80KrcCHST9mzBXPwwRdpZYVx8kaQRwdal/ErBQceoaQay5QqSFvRF4EHiD5jkdOFXSMCbXYPsRAnhkaXeXyrErgFdt/3s6rusvwLckPQ8cUcaHpCUk3Vyu8RFCuAwlTOjtCPM9wBXFzD6K0DhPKuUrAU9KeoZIu1vTsDsDd0saSZjK77B9Yzm2GrE8MKN0Ae4pywm1ZQcIIX5I6ftBvprO93ygT7mvK/KlxaUHMKI8i52AvlPpY6rYfgq4BvgX8X/qoNoyjqSbJS1RPh8i6TVCUx+pCP87tXbHEFahx8rfiZXlJQgLRAr8pKHI0LpJ0koknQcMs/2Xeo+lNUi6zfYU96UnSTL7kQI/SVqBYn/1OMLLfEK9x5MkSTItUuAnSZIkSRsg1/CTJEmSpA2QAj9JkiRJ2gAp8JMkSZKkDZACP0mSJEnaACnwkyRJkqQNkAI/SZIkSdoA/w/GqoMU6qKjmQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "target_names = [\"Negative(n)\", \"Neutral(o)\", \"Positive(p)\"]\n",
    "plot_confusion_matrix(matrix,\n",
    "                          target_names,\n",
    "                          title='Confusion matrix for tfidf',\n",
    "                          cmap=None,\n",
    "                          normalize=True)"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}