Rare Occupation Classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Rare Occupation Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This example notebook uses hypothetical data to show how the `BinaryContextTransformer` could be used to help classify occupation codes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy as sp\n",
    "\n",
    "from binarycontexttransformer import BinaryContextTransformer\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from IPython.display import display"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These helper functions will provide visualizations of the data, predictions, and model weights."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_preds(df, X, y, clf, occupations, duplicates=False):\n",
    "    df = df.copy()\n",
    "    df[\"pred\"] = [occupations[code] for code in clf.predict(X)]\n",
    "    df[\"proba\"] = clf.predict_proba(X).max(axis=1)\n",
    "    df[\"correct\"] = df[\"pred\"] == df[\"occupation\"]\n",
    "    if not duplicates:\n",
    "        df = df.drop_duplicates()\n",
    "    acc = accuracy_score(df[\"occupation\"], df[\"pred\"])\n",
    "    pred_df = df[[\"correct\", \"occupation\", \"pred\", \"proba\", \"jobtitle\", \"industry\"]]\n",
    "    out = pred_df.style.apply(lambda row: [\n",
    "        \"background: mediumspringgreen;\" if row[\"pred\"] is row[\"occupation\"] else \"background: tomato;\" for cell in row\n",
    "    ], axis=1, subset=[\"occupation\", \"pred\"])\n",
    "    print(\"Accuracy = {:.3f}\".format(acc))\n",
    "    return out\n",
    "\n",
    "\n",
    "def show_weights(clf, features, occupations, ax=None):\n",
    "    has_plot = True\n",
    "    if ax is None:\n",
    "        _, ax = plt.subplots(1, 1)\n",
    "        has_plot = False\n",
    "    weights = np.hstack([clf.intercept_.reshape(-1, 1), clf.coef_]).T\n",
    "    classes = [occupations[code] for code in clf.classes_]\n",
    "    predictors = [\"intercept\"] + features\n",
    "    class_labels = [\"\\n\".join(c.split(\" \")) for c in classes]\n",
    "    sns.heatmap(weights, annot=True, fmt=\".3f\", cmap=\"RdBu\", ax=ax)\n",
    "    ax.set_xticks(np.arange(len(classes)) + 0.5)\n",
    "    ax.set_yticks(np.arange(len(predictors)) + 0.5)\n",
    "    ax.set_xticklabels(class_labels, {\"rotation\": 0})\n",
    "    ax.set_yticklabels(predictors, {\"rotation\": 0})\n",
    "    ax.set_xlabel(\"Target Class\")\n",
    "    ax.set_ylabel(\"Predictor Variable\")\n",
    "    ax.set_title(\"Model Weights\")\n",
    "    if not has_plot:\n",
    "        plt.show()\n",
    "\n",
    "    \n",
    "def show_matrix(X, labels, features, occupations, duplicates=False):\n",
    "    out = pd.DataFrame(X.todense(), columns=features, index=labels)\n",
    "    if not duplicates:\n",
    "        out = out.drop_duplicates()\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "fields = [\"occupation\", \"code\", \"jobtitle\", \"industry\"]\n",
    "\n",
    "occupations = {\n",
    "    \"27-3021\": \"Broadcast News Analysts\",\n",
    "    \"13-2051\": \"Financial Analysts\",\n",
    "    \"15-1121\": \"Computer Systems Analysts\",\n",
    "    \"13-2011\": \"Accountants and Auditors\",\n",
    "    \"15-1132\": \"Software Developers, Applications\",\n",
    "    \"27-3022\": \"Reporters and Correspondents\"\n",
    "}\n",
    "\n",
    "data = [\n",
    "    (\"13-2051\", \"financial analyst\", \"finance\"),\n",
    "    (\"13-2051\", \"financial analyst\", \"finance\"),\n",
    "    (\"27-3021\", \"news analyst\", \"news\"),\n",
    "    (\"27-3021\", \"news analyst\", \"news\"),\n",
    "    (\"15-1121\", \"computer analyst\", \"tech\"),\n",
    "    (\"15-1121\", \"computer analyst\", \"tech\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"13-2011\", \"accountant\", \"finance\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"27-3022\", \"reporter\", \"news\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "    (\"15-1132\", \"developer\", \"tech\"),\n",
    "]\n",
    "\n",
    "data_validate = [\n",
    "    (\"15-1121\", \"analyst\", \"tech\"),\n",
    "    (\"27-3021\", \"analyst\", \"news\"),\n",
    "    (\"13-2051\", \"analyst\", \"finance\"),\n",
    "]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For this hypothetical example, the task is to predict what kind of occupation a job falls under, based on the job title and the industry.\n",
    "\n",
    "In this hypothetical universe, let us assume that there are only six different occupations and only three different industries. In each industry, there are two occupations: one that occurs more frequently in the training data and one that occurs relatively less frequently.\n",
    "\n",
    "The table shows the six distinct job titles, the corresponding occupation classifications, and the frequencies in the training data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>occupation</th>\n",
       "      <th>code</th>\n",
       "      <th>jobtitle</th>\n",
       "      <th>industry</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Accountants and Auditors</td>\n",
       "      <td>13-2011</td>\n",
       "      <td>accountant</td>\n",
       "      <td>finance</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Broadcast News Analysts</td>\n",
       "      <td>27-3021</td>\n",
       "      <td>news analyst</td>\n",
       "      <td>news</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Computer Systems Analysts</td>\n",
       "      <td>15-1121</td>\n",
       "      <td>computer analyst</td>\n",
       "      <td>tech</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Financial Analysts</td>\n",
       "      <td>13-2051</td>\n",
       "      <td>financial analyst</td>\n",
       "      <td>finance</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Reporters and Correspondents</td>\n",
       "      <td>27-3022</td>\n",
       "      <td>reporter</td>\n",
       "      <td>news</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Software Developers, Applications</td>\n",
       "      <td>15-1132</td>\n",
       "      <td>developer</td>\n",
       "      <td>tech</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          occupation     code           jobtitle industry  \\\n",
       "0           Accountants and Auditors  13-2011         accountant  finance   \n",
       "1            Broadcast News Analysts  27-3021       news analyst     news   \n",
       "2          Computer Systems Analysts  15-1121   computer analyst     tech   \n",
       "3                 Financial Analysts  13-2051  financial analyst  finance   \n",
       "4       Reporters and Correspondents  27-3022           reporter     news   \n",
       "5  Software Developers, Applications  15-1132          developer     tech   \n",
       "\n",
       "  count  \n",
       "0     7  \n",
       "1     2  \n",
       "2     2  \n",
       "3     2  \n",
       "4     7  \n",
       "5     7  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame(data, columns=[\"code\", \"jobtitle\", \"industry\"])\n",
    "df[\"occupation\"] = df[\"code\"].apply(lambda code: occupations[code])\n",
    "pd.DataFrame(df.groupby(fields).size().reset_index().values, columns=fields + [\"count\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Suppose that the validation data, which is not used to train a model, contains job titles that are less informative compared to the training data. In the data below, every record has the job title `\"analyst\"`, but each is a different kind of occupation. Using information about the industry, a human should be able to figure out which kind of analyst is the correct match.\n",
    "\n",
    "_**Note:** It might be more reasonable to assume that some data like this will be part of the training dataset. However, in large datasets, if these low-information titles are infrequent in the training data, a model might not learn appropriate weights for each word. Or, if the low-information titles are very frequent in the training data, a model might learn the wrong associations between words and occupations._"
   ]
  },
  {
   "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>occupation</th>\n",
       "      <th>code</th>\n",
       "      <th>jobtitle</th>\n",
       "      <th>industry</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Computer Systems Analysts</td>\n",
       "      <td>15-1121</td>\n",
       "      <td>analyst</td>\n",
       "      <td>tech</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Broadcast News Analysts</td>\n",
       "      <td>27-3021</td>\n",
       "      <td>analyst</td>\n",
       "      <td>news</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Financial Analysts</td>\n",
       "      <td>13-2051</td>\n",
       "      <td>analyst</td>\n",
       "      <td>finance</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  occupation     code jobtitle industry\n",
       "0  Computer Systems Analysts  15-1121  analyst     tech\n",
       "1    Broadcast News Analysts  27-3021  analyst     news\n",
       "2         Financial Analysts  13-2051  analyst  finance"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_validate = pd.DataFrame(data_validate, columns=[\"code\", \"jobtitle\", \"industry\"])\n",
    "df_validate[\"occupation\"] = df_validate[\"code\"].apply(lambda code: occupations[code])\n",
    "df_validate[fields]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Basic Model\n",
    "\n",
    "Let us try to train a logistic regression model (with L2 penalty) to predict the occupation code. This model uses binary variables for each job title word and each industry as predictors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "jobtitle_vectorizer = CountVectorizer(analyzer=\"word\", binary=True)\n",
    "industry_vectorizer = CountVectorizer(analyzer=\"word\", binary=True)\n",
    "X_jobtitle = jobtitle_vectorizer.fit_transform(df[\"jobtitle\"])\n",
    "X_industry = industry_vectorizer.fit_transform(df[\"industry\"])\n",
    "X_basic = sp.sparse.hstack([X_jobtitle, X_industry])\n",
    "y = df[\"code\"]\n",
    "\n",
    "features_jobtitle = [\"job_\" + f for f in jobtitle_vectorizer.get_feature_names()]\n",
    "features_industry = [\"ind_\" + f for f in industry_vectorizer.get_feature_names()]\n",
    "features_basic = features_jobtitle + features_industry\n",
    "\n",
    "clf_basic = LogisticRegression(solver=\"lbfgs\", multi_class=\"multinomial\", fit_intercept=True, penalty=\"l2\")\n",
    "clf_basic.fit(X_basic, y)\n",
    "y_basic = clf_basic.predict(X_basic)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.1. Basic Model Matrix\n",
    "\n",
    "This table shows how each of the six distinct job titles in our training data gets represented in the input data. There are 10 total features, 7 from job title words and 3 from industries."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10 features in `term` matrix\n",
      "7 features from `jobtitle`\n",
      "3 features from `industry`\n"
     ]
    },
    {
     "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>job_accountant</th>\n",
       "      <th>job_analyst</th>\n",
       "      <th>job_computer</th>\n",
       "      <th>job_developer</th>\n",
       "      <th>job_financial</th>\n",
       "      <th>job_news</th>\n",
       "      <th>job_reporter</th>\n",
       "      <th>ind_finance</th>\n",
       "      <th>ind_news</th>\n",
       "      <th>ind_tech</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>financial analyst</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>news analyst</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>computer analyst</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>accountant</th>\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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>reporter</th>\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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>developer</th>\n",
       "      <td>0</td>\n",
       "      <td>0</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>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   job_accountant  job_analyst  job_computer  job_developer  \\\n",
       "financial analyst               0            1             0              0   \n",
       "news analyst                    0            1             0              0   \n",
       "computer analyst                0            1             1              0   \n",
       "accountant                      1            0             0              0   \n",
       "reporter                        0            0             0              0   \n",
       "developer                       0            0             0              1   \n",
       "\n",
       "                   job_financial  job_news  job_reporter  ind_finance  \\\n",
       "financial analyst              1         0             0            1   \n",
       "news analyst                   0         1             0            0   \n",
       "computer analyst               0         0             0            0   \n",
       "accountant                     0         0             0            1   \n",
       "reporter                       0         0             1            0   \n",
       "developer                      0         0             0            0   \n",
       "\n",
       "                   ind_news  ind_tech  \n",
       "financial analyst         0         0  \n",
       "news analyst              1         0  \n",
       "computer analyst          0         1  \n",
       "accountant                0         0  \n",
       "reporter                  1         0  \n",
       "developer                 0         1  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"{} features in `term` matrix\".format(X_basic.shape[1]))\n",
    "print(\"{} features from `jobtitle`\".format(X_jobtitle.shape[1]))\n",
    "print(\"{} features from `industry`\".format(X_industry.shape[1]))\n",
    "show_matrix(X_basic, df[\"jobtitle\"].values, features_basic, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.2. Basic Model Predictions\n",
    "\n",
    "With this model, we can achieve 100% accuracy on the training dataset! Hooray!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy = 1.000\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }</style><table id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >correct</th>        <th class=\"col_heading level0 col1\" >occupation</th>        <th class=\"col_heading level0 col2\" >pred</th>        <th class=\"col_heading level0 col3\" >proba</th>        <th class=\"col_heading level0 col4\" >jobtitle</th>        <th class=\"col_heading level0 col5\" >industry</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col0\" class=\"data row0 col0\" >True</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col1\" class=\"data row0 col1\" >Financial Analysts</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col2\" class=\"data row0 col2\" >Financial Analysts</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col3\" class=\"data row0 col3\" >0.545106</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col4\" class=\"data row0 col4\" >financial analyst</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row0_col5\" class=\"data row0 col5\" >finance</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18level0_row1\" class=\"row_heading level0 row1\" >2</th>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col0\" class=\"data row1 col0\" >True</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col1\" class=\"data row1 col1\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col2\" class=\"data row1 col2\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col3\" class=\"data row1 col3\" >0.545106</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col4\" class=\"data row1 col4\" >news analyst</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row1_col5\" class=\"data row1 col5\" >news</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18level0_row2\" class=\"row_heading level0 row2\" >4</th>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col0\" class=\"data row2 col0\" >True</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col1\" class=\"data row2 col1\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col2\" class=\"data row2 col2\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col3\" class=\"data row2 col3\" >0.545106</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col4\" class=\"data row2 col4\" >computer analyst</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row2_col5\" class=\"data row2 col5\" >tech</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18level0_row3\" class=\"row_heading level0 row3\" >6</th>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col0\" class=\"data row3 col0\" >True</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col1\" class=\"data row3 col1\" >Accountants and Auditors</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col2\" class=\"data row3 col2\" >Accountants and Auditors</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col3\" class=\"data row3 col3\" >0.822282</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col4\" class=\"data row3 col4\" >accountant</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row3_col5\" class=\"data row3 col5\" >finance</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18level0_row4\" class=\"row_heading level0 row4\" >13</th>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col0\" class=\"data row4 col0\" >True</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col1\" class=\"data row4 col1\" >Reporters and Correspondents</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col2\" class=\"data row4 col2\" >Reporters and Correspondents</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col3\" class=\"data row4 col3\" >0.822282</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col4\" class=\"data row4 col4\" >reporter</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row4_col5\" class=\"data row4 col5\" >news</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18level0_row5\" class=\"row_heading level0 row5\" >20</th>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col0\" class=\"data row5 col0\" >True</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col1\" class=\"data row5 col1\" >Software Developers, Applications</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col2\" class=\"data row5 col2\" >Software Developers, Applications</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col3\" class=\"data row5 col3\" >0.822282</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col4\" class=\"data row5 col4\" >developer</td>\n",
       "                        <td id=\"T_95ce4300_e58c_11e9_97e3_f40f2436fa18row5_col5\" class=\"data row5 col5\" >tech</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x114181dd8>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_preds(df, X_basic, y, clf_basic, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "However, on the validation dataset, the accuracy is 0%..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy = 0.000\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col1 {\n",
       "            background:  tomato;\n",
       "        }    #T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col2 {\n",
       "            background:  tomato;\n",
       "        }    #T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col1 {\n",
       "            background:  tomato;\n",
       "        }    #T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col2 {\n",
       "            background:  tomato;\n",
       "        }    #T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col1 {\n",
       "            background:  tomato;\n",
       "        }    #T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col2 {\n",
       "            background:  tomato;\n",
       "        }</style><table id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >correct</th>        <th class=\"col_heading level0 col1\" >occupation</th>        <th class=\"col_heading level0 col2\" >pred</th>        <th class=\"col_heading level0 col3\" >proba</th>        <th class=\"col_heading level0 col4\" >jobtitle</th>        <th class=\"col_heading level0 col5\" >industry</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col0\" class=\"data row0 col0\" >False</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col1\" class=\"data row0 col1\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col2\" class=\"data row0 col2\" >Software Developers, Applications</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col3\" class=\"data row0 col3\" >0.325774</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col4\" class=\"data row0 col4\" >analyst</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row0_col5\" class=\"data row0 col5\" >tech</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col0\" class=\"data row1 col0\" >False</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col1\" class=\"data row1 col1\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col2\" class=\"data row1 col2\" >Reporters and Correspondents</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col3\" class=\"data row1 col3\" >0.325774</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col4\" class=\"data row1 col4\" >analyst</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row1_col5\" class=\"data row1 col5\" >news</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col0\" class=\"data row2 col0\" >False</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col1\" class=\"data row2 col1\" >Financial Analysts</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col2\" class=\"data row2 col2\" >Accountants and Auditors</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col3\" class=\"data row2 col3\" >0.325774</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col4\" class=\"data row2 col4\" >analyst</td>\n",
       "                        <td id=\"T_95d6da42_e58c_11e9_9b44_f40f2436fa18row2_col5\" class=\"data row2 col5\" >finance</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x117396eb8>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_jobtitle_validate = jobtitle_vectorizer.transform(df_validate[\"jobtitle\"])\n",
    "X_industry_validate = industry_vectorizer.transform(df_validate[\"industry\"])\n",
    "X_basic_validate = sp.sparse.hstack([\n",
    "    X_jobtitle_validate,\n",
    "    X_industry_validate,\n",
    "])\n",
    "show_preds(df_validate, X_basic_validate, y, clf_basic, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Notice that when the model misclassifies each analyst, it does not predict a different kind of analyst. In fact, it predicts a different kind of occupation altogether. This seems strange.\n",
    "\n",
    "1. Why does the model classify the job as a different occupation, instead of a different kind of analyst?\n",
    "2. Why does the model fail to distinguish different types of analysts when it has industry as a predictor variable?\n",
    "\n",
    "The answer involves the multiclass nature of the task and imbalances in the frequencies of the occupations.\n",
    "\n",
    "- The word `\"analyst\"` is not predictive in this case because in the training data that word appears in three different occupations at the same frequency in each occupation. Each of the three analyst occupations is equally likely.\n",
    "- The model learns a weight between each industry and each occupation. The larger weights correspond to the occupations that are more frequent in that industry.\n",
    "\n",
    "Thus, the model ends up predicting the occupation that is most common in the given industry, rather than a different kind of analyst."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.3. Basic Model Weights\n",
    "\n",
    "To understand this outcome, it may be helpful to inspect the weights of the logistic regression model. Each row in the grid below represents a predictor variable and each column represents an occupation.\n",
    "\n",
    "Take for example an analyst working in the news industry. The model misclassifies their occupation as `\"Reporters and Correspondents\"` when they should fall under `\"Broadcast News Analysts\"`.\n",
    "\n",
    "As expected, the highest coefficients for the word `\"analyst\"` are in the three analyst occupations. These coefficients and the intercepts are all equal. This means that a prediction based only on the word `\"analyst\"` would given equal predicted probabilities for each of those three occupations.\n",
    "\n",
    "The industry variable breaks the tie. Since the record is in the news industry, only two weights activate in the industry section of this grid. The news industry coefficient weight for `\"Reporters and Correspondents\"` is higher than the weight for `\"Broadcast News Analysts\"`, so the former is the prediction output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117396320>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_weights(clf_basic, features_basic, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Context Model\n",
    "\n",
    "The `BinaryContextTransformer` may be useful in creating a more detailed representation of the input data: allowing us to learn different weights for each job title word, based on the industry it appears in. We will observe how this changes the model weights and the predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Help on class BinaryContextTransformer in module binarycontexttransformer:\n",
      "\n",
      "class BinaryContextTransformer(sklearn.base.TransformerMixin)\n",
      " |  Expands base features into interaction terms when they appear with\n",
      " |  different context features. Base features are variables that may have different\n",
      " |  meanings in different contexts. Context features are indicator variables that\n",
      " |  denote which context a record belongs to. Both base features and context features\n",
      " |  must be binary.\n",
      " |  \n",
      " |  Method resolution order:\n",
      " |      BinaryContextTransformer\n",
      " |      sklearn.base.TransformerMixin\n",
      " |      builtins.object\n",
      " |  \n",
      " |  Methods defined here:\n",
      " |  \n",
      " |  __init__(self, features, contexts, progress=None)\n",
      " |      Args:\n",
      " |          features: names of base feature columns for input matrix\n",
      " |          contexts: names of context feature columns for input matrix\n",
      " |          progress: function of format progress_fn(iter, total) that takes\n",
      " |              an iterable and an integer with the total number of items and\n",
      " |              returns a generator to track progress at each step of the\n",
      " |              iterable (default=None)\n",
      " |  \n",
      " |  fit(self, X, X_context)\n",
      " |      Args:\n",
      " |          X: input matrix, base feature columns\n",
      " |          X_context: input matrix, context feature columns\n",
      " |  \n",
      " |  fit_transform(self, X, X_context)\n",
      " |      Args:\n",
      " |          X: input matrix, base feature columns\n",
      " |          X_context: input matrix, context feature columns\n",
      " |  \n",
      " |  get_feature_names(self)\n",
      " |      Returns a list of feature names corresponding to column indices.\n",
      " |  \n",
      " |  transform(self, X, X_context)\n",
      " |      Args:\n",
      " |          X: input matrix, base feature columns\n",
      " |          X_context: input matrix, context feature columns\n",
      " |  \n",
      " |  ----------------------------------------------------------------------\n",
      " |  Data descriptors inherited from sklearn.base.TransformerMixin:\n",
      " |  \n",
      " |  __dict__\n",
      " |      dictionary for instance variables (if defined)\n",
      " |  \n",
      " |  __weakref__\n",
      " |      list of weak references to the object (if defined)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "help(BinaryContextTransformer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `BinaryContextTransformer` takes the matrices produced by the `CountVectorizer` transformers as input. The job title words are the **base** features and the industries are the **context** features. Each record is part of only one industry. We stack the two input matrices from the `CountVectorizer` transformers and the matrix output by the `BinaryContextTransformer` and use this as the input to our new model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "bct = BinaryContextTransformer(\n",
    "    features=features_jobtitle,\n",
    "    contexts=features_industry\n",
    ")\n",
    "X_jobtitle_industry = bct.fit_transform(X_jobtitle, X_industry)\n",
    "X_context = sp.sparse.hstack([X_jobtitle, X_industry, X_jobtitle_industry])\n",
    "features_context = features_basic + bct.get_feature_names()\n",
    "\n",
    "clf_context = LogisticRegression(solver=\"lbfgs\", multi_class=\"multinomial\", fit_intercept=True, penalty=\"l2\")\n",
    "clf_context.fit(X_context, y)\n",
    "y_context = clf_context.predict(X_context)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.1. Context Model Matrix\n",
    "\n",
    "For an alternative model, we still use a logistic regression (with L2 penalty). This time, we add the binary context terms for interactions between job title words and industry. Now there are 13 total features.\n",
    "\n",
    "Notice that we only added binary context features for job title words that appear in at least two different industries. The only word that appears in multiple industries is `\"analyst\"` and it appears in all three industries."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13 features in `context` matrix\n",
      "7 features from `jobtitle`\n",
      "3 features from `industry`\n",
      "3 features from `jobtitle_x_industry`\n"
     ]
    },
    {
     "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>job_accountant</th>\n",
       "      <th>job_analyst</th>\n",
       "      <th>job_computer</th>\n",
       "      <th>job_developer</th>\n",
       "      <th>job_financial</th>\n",
       "      <th>job_news</th>\n",
       "      <th>job_reporter</th>\n",
       "      <th>ind_finance</th>\n",
       "      <th>ind_news</th>\n",
       "      <th>ind_tech</th>\n",
       "      <th>ind_finance_x_job_analyst</th>\n",
       "      <th>ind_news_x_job_analyst</th>\n",
       "      <th>ind_tech_x_job_analyst</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>financial analyst</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>news analyst</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>computer analyst</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>accountant</th>\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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>reporter</th>\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>1</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>developer</th>\n",
       "      <td>0</td>\n",
       "      <td>0</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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   job_accountant  job_analyst  job_computer  job_developer  \\\n",
       "financial analyst               0            1             0              0   \n",
       "news analyst                    0            1             0              0   \n",
       "computer analyst                0            1             1              0   \n",
       "accountant                      1            0             0              0   \n",
       "reporter                        0            0             0              0   \n",
       "developer                       0            0             0              1   \n",
       "\n",
       "                   job_financial  job_news  job_reporter  ind_finance  \\\n",
       "financial analyst              1         0             0            1   \n",
       "news analyst                   0         1             0            0   \n",
       "computer analyst               0         0             0            0   \n",
       "accountant                     0         0             0            1   \n",
       "reporter                       0         0             1            0   \n",
       "developer                      0         0             0            0   \n",
       "\n",
       "                   ind_news  ind_tech  ind_finance_x_job_analyst  \\\n",
       "financial analyst         0         0                          1   \n",
       "news analyst              1         0                          0   \n",
       "computer analyst          0         1                          0   \n",
       "accountant                0         0                          0   \n",
       "reporter                  1         0                          0   \n",
       "developer                 0         1                          0   \n",
       "\n",
       "                   ind_news_x_job_analyst  ind_tech_x_job_analyst  \n",
       "financial analyst                       0                       0  \n",
       "news analyst                            1                       0  \n",
       "computer analyst                        0                       1  \n",
       "accountant                              0                       0  \n",
       "reporter                                0                       0  \n",
       "developer                               0                       0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"{} features in `context` matrix\".format(X_context.shape[1]))\n",
    "print(\"{} features from `jobtitle`\".format(X_jobtitle.shape[1]))\n",
    "print(\"{} features from `industry`\".format(X_industry.shape[1]))\n",
    "print(\"{} features from `jobtitle_x_industry`\".format(X_jobtitle_industry.shape[1]))\n",
    "show_matrix(X_context, df[\"jobtitle\"].values, features_context, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2. Context Model Predictions\n",
    "\n",
    "Just as with the basic model, our context model achieves 100% accuracy on the training set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy = 1.000\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }</style><table id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >correct</th>        <th class=\"col_heading level0 col1\" >occupation</th>        <th class=\"col_heading level0 col2\" >pred</th>        <th class=\"col_heading level0 col3\" >proba</th>        <th class=\"col_heading level0 col4\" >jobtitle</th>        <th class=\"col_heading level0 col5\" >industry</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col0\" class=\"data row0 col0\" >True</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col1\" class=\"data row0 col1\" >Financial Analysts</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col2\" class=\"data row0 col2\" >Financial Analysts</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col3\" class=\"data row0 col3\" >0.628344</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col4\" class=\"data row0 col4\" >financial analyst</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row0_col5\" class=\"data row0 col5\" >finance</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18level0_row1\" class=\"row_heading level0 row1\" >2</th>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col0\" class=\"data row1 col0\" >True</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col1\" class=\"data row1 col1\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col2\" class=\"data row1 col2\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col3\" class=\"data row1 col3\" >0.628344</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col4\" class=\"data row1 col4\" >news analyst</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row1_col5\" class=\"data row1 col5\" >news</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18level0_row2\" class=\"row_heading level0 row2\" >4</th>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col0\" class=\"data row2 col0\" >True</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col1\" class=\"data row2 col1\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col2\" class=\"data row2 col2\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col3\" class=\"data row2 col3\" >0.628344</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col4\" class=\"data row2 col4\" >computer analyst</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row2_col5\" class=\"data row2 col5\" >tech</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18level0_row3\" class=\"row_heading level0 row3\" >6</th>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col0\" class=\"data row3 col0\" >True</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col1\" class=\"data row3 col1\" >Accountants and Auditors</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col2\" class=\"data row3 col2\" >Accountants and Auditors</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col3\" class=\"data row3 col3\" >0.830019</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col4\" class=\"data row3 col4\" >accountant</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row3_col5\" class=\"data row3 col5\" >finance</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18level0_row4\" class=\"row_heading level0 row4\" >13</th>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col0\" class=\"data row4 col0\" >True</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col1\" class=\"data row4 col1\" >Reporters and Correspondents</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col2\" class=\"data row4 col2\" >Reporters and Correspondents</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col3\" class=\"data row4 col3\" >0.830019</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col4\" class=\"data row4 col4\" >reporter</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row4_col5\" class=\"data row4 col5\" >news</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18level0_row5\" class=\"row_heading level0 row5\" >20</th>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col0\" class=\"data row5 col0\" >True</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col1\" class=\"data row5 col1\" >Software Developers, Applications</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col2\" class=\"data row5 col2\" >Software Developers, Applications</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col3\" class=\"data row5 col3\" >0.830019</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col4\" class=\"data row5 col4\" >developer</td>\n",
       "                        <td id=\"T_963c3718_e58c_11e9_ad0c_f40f2436fa18row5_col5\" class=\"data row5 col5\" >tech</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x11803e828>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_preds(df, X_context, y, clf_context, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This time, the context model also achieves 100% accuracy on the validation set!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy = 1.000\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_96439986_e58c_11e9_a171_f40f2436fa18row0_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_96439986_e58c_11e9_a171_f40f2436fa18row0_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_96439986_e58c_11e9_a171_f40f2436fa18row1_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_96439986_e58c_11e9_a171_f40f2436fa18row1_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_96439986_e58c_11e9_a171_f40f2436fa18row2_col1 {\n",
       "            background:  mediumspringgreen;\n",
       "        }    #T_96439986_e58c_11e9_a171_f40f2436fa18row2_col2 {\n",
       "            background:  mediumspringgreen;\n",
       "        }</style><table id=\"T_96439986_e58c_11e9_a171_f40f2436fa18\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >correct</th>        <th class=\"col_heading level0 col1\" >occupation</th>        <th class=\"col_heading level0 col2\" >pred</th>        <th class=\"col_heading level0 col3\" >proba</th>        <th class=\"col_heading level0 col4\" >jobtitle</th>        <th class=\"col_heading level0 col5\" >industry</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_96439986_e58c_11e9_a171_f40f2436fa18level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row0_col0\" class=\"data row0 col0\" >True</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row0_col1\" class=\"data row0 col1\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row0_col2\" class=\"data row0 col2\" >Computer Systems Analysts</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row0_col3\" class=\"data row0 col3\" >0.399234</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row0_col4\" class=\"data row0 col4\" >analyst</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row0_col5\" class=\"data row0 col5\" >tech</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_96439986_e58c_11e9_a171_f40f2436fa18level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row1_col0\" class=\"data row1 col0\" >True</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row1_col1\" class=\"data row1 col1\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row1_col2\" class=\"data row1 col2\" >Broadcast News Analysts</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row1_col3\" class=\"data row1 col3\" >0.399234</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row1_col4\" class=\"data row1 col4\" >analyst</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row1_col5\" class=\"data row1 col5\" >news</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_96439986_e58c_11e9_a171_f40f2436fa18level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row2_col0\" class=\"data row2 col0\" >True</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row2_col1\" class=\"data row2 col1\" >Financial Analysts</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row2_col2\" class=\"data row2 col2\" >Financial Analysts</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row2_col3\" class=\"data row2 col3\" >0.399234</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row2_col4\" class=\"data row2 col4\" >analyst</td>\n",
       "                        <td id=\"T_96439986_e58c_11e9_a171_f40f2436fa18row2_col5\" class=\"data row2 col5\" >finance</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x118150a90>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_jobtitle_validate = jobtitle_vectorizer.transform(df_validate[\"jobtitle\"])\n",
    "X_industry_validate = industry_vectorizer.transform(df_validate[\"industry\"])\n",
    "X_context_validate = sp.sparse.hstack([\n",
    "    X_jobtitle_validate,\n",
    "    X_industry_validate,\n",
    "    bct.transform(X_jobtitle_validate, X_industry_validate)\n",
    "])\n",
    "show_preds(df_validate, X_context_validate, y, clf_context, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What happened? Could this be model overfitting?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3. Context Model Weights\n",
    "\n",
    "Just as in the basic model, the coefficients for the word `\"analyst\"` and the three analyst occupations have equal values. The industry coefficients have higher values for the more frequent occupation in their industry.\n",
    "\n",
    "Now, there are additional coefficients for the binary context features. These coefficients have higher values for the binary context features between each industry and the word `\"analyst\"`, which pushes the prediction output to the correct answer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAewAAAEuCAYAAABMCiHfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAIABJREFUeJzsnXd8FNUWx79nN5veGwmB0KUK0kEUQRQVRWyoqCiIig27PssTLA8s+Kwoz45iF7EDFjqIdELHQIBQAqmkkbbZ+/6YSS9syEay8X4/n/1k986ZO2duZufsLXN+opRCo9FoNBpN48Zyqh3QaDQajUZzYnTA1mg0Go3GDdABW6PRaDQaN0AHbI1Go9Fo3AAdsDUajUajcQN0wNZoNBqNxg3QAVuj0dQZEWktIkpEPJywHSciK/4mv7aJyBAnbfeJyHkN7JJG4zJ0wNZomjhmYCoUkfBK5ZvMoNv61HgGIvKLiDxS7nOM6VN1ZVEnqk8p1VUptcQFfg0RkYP1rUejcSU6YGs0/wz2AmNKPojI6YDPqXOnlGXAOeU+DwZ2VlMWr5Q68nc6ptE0NnTA1mj+GcwGbiz3+Sbg4/IGIhIkIh+LSIqI7BeRf4uIxdxmFZGXRCRVRBKAi6vZ930RSRKRQyLyHxGxOuHXMmBQyXGAs4FXgT6VypaVO9Yl5ujAMRH5Q0S6l9tWOswtIj4i8pGIZIjIDhF5pJpe8xkisllEMkXkSxHxFhE/YD7QXERyzFdzEeknIutEJEtEjorIy06cn0bjMnTA1mj+GfwJBIpIZzOQXgN8UsnmDSAIaIvRw70RGG9uuxW4BOgJ9AGuqrTvR4AdaG/aDAduccKvNYAX0MP8PBj4DdhdqWwZgIj0Aj4AJgJhwNvADyLiVU3dU4DW5vmcD9xQjc3VwIVAG6A7ME4plQtcBBxWSvmbr8PAa8BrSqlAoB3wlRPnp9G4DB2wNZp/DiW97PMxhp0PlWwoF8QfU0plK6X2Af8FxpomVwOvKqUOKKXSgefK7dsMI8Ddp5TKVUolA68A157IIaVUAbAaGCwioUCwUioBWF6urAuw1NzlVuBtpdRqpVSxUuojoAAYUE31VwPTlFIZSqmDwOvV2LyulDpsntOPwBm1uFsEtBeRcKVUjlLqzxOdn0bjSk64wlOj0TQZZmP0VNtQaTgcCAc8gf3lyvYDMeb75sCBSttKaAXYgCQRKSmzVLKvjWUYveh9QMlq8hUYvft9wAGlVMnxWgE3icikcvt7mv5VprLP1flTfl78eA31lDABeAbYKSJ7gaeVUj/VYq/RuBQdsDWafwhKqf1moBmBEXzKk4rRg2wFbDfLYinrhScBLcvZx5Z7fwCjlxuulLKfhGvLgNsxgvNys2wl8J5Ztqyc7QFgqlJqqhP1JgEtKDuflrXYVqaKjKFSKh4YY86tXwHMEZEwcwhdo2lw9JC4RvPPYgJwbuUgo5QqxpiTnSoiASLSCniAsnnur4B7RKSFiIQAj5bbNwn4FfiviASKiEVE2olI+ZXetfEHEIwxx7zcrDMDSDHLygfsd4HbRaS/GPiJyMUiElBNvV8Bj4lIiIjEAHc76Q/AUSBMRIJKCkTkBhGJUEo5gGNmcXEd6tRo6oUO2BrNPwil1B6l1LoaNk8CcoEEjCHpzzAWeIERKH8B4oANwNxK+96IMTS9HcgA5gDRTvp0HFiPsfhsa7lNy4FIygVs0/dbgRnmcXYD42qo+hngIMYjbb+bPhU46dNO4HMgwVyN3hxjcdo2EcnBWIB2rVIq35n6NBpXIEpVGfnRaDSaJoeI3IERZJ3t+Ws0jQrdw9ZoNE0SEYkWkUHmEH1H4EHg21Ptl0ZzsuhFZxqNpqniifGcdhuMOecvgLdOqUcaTT3QQ+IajUaj0bgBekhco9FoNBo3QAdsjUaj0WjcAD2HrTklvLx8j1vNxfw1eNipduEfwWnLFp5qF+qEO14X7tbGAA+c3U5ObFU7nj1vdvqeU7jxg3ofryHQPWyNRqPRaNwA3cPWaDQaTZPH4uF5ql2oNzpgaxo1iVvX8cfnb6McDjqdfQE9R1xdxWbP2mWs++FTRISwFm0Ydtu/SrcV5h3nyycn0qbnQM66/k4Aiu1FrPhsJkm7NiNioe/lN9K291ku9fvq16bQbcRQCo/n8dG4hziwcVsVG6vNxrUznua0IQNQDsX3T0xn49wFpdt7XXkRt82ZybQ+I0lcvwWAmNM7cf3b0/AO9Ec5HDzXdxT2AqeSdzUpf/V10fDt7K5tXBNicUaevXGjA7YbISJ/KKXOPIHNfcA7ZrrHvx1XHt/hKGblp29x8QNT8QsJZ+5/7qP1GQMIaV6mO5F59BAb533FZY++hJdfAHlZxyrUsfa7j2l+WrcKZRt+/hKfgCCunfoeyuEgPze7vq5WoNtFQ4js0IbJHYbQpn9Prps5lRcGXFbF7qIn7iY7OY0pHc9FRPANDS7d5uXvx9B7xpHw58bSMovVyvhPXuHDsQ9waPMO/EKDKS4q+sf5q6+Lhm9nd23j2mgKAVvPYbsRJwrWJvcBvnWp19RCdhV1Pn5NJO/9i8DI5gRGRGP1sNG+32D2bVpVwWbHsgV0HXoJXn6G9oNPYNnNLWVfPHlZx2jRtVeFfXat+JWeI64BQCwWfAKCcCXdRw3nz4+NVNt7V2/EJziAwKiIKnZn3jyaBc8ZeTyUUuSmZZRuu/TZB/n1xbex55f1kroMP5tDm3dyaPMOAHLTj6Ecjn+cv/q6aPh2dtc2rg2xWp1+NVZ0wHYjTNEBRGSIiCwRkTkislNEPjWVi+7B0PNdLCKLTdvhIrJKRDaIyNci4m+W7xORySKyAhgtIu1F5HcRiTNt25l2D4vIWhHZLCJPm2WtzeN+ZJbPERHf6o5fH45npOEfEl762S8knNyMtAo2mUcPkXn0EN899yDfTrufxK2GroVyOFj11XsMGF1RRbLgeA5g/Pr/5plJ/DZzGsczM3AlwTHNyDhwuPTzsYNHCI6JqmDjExQIGDfgx9f/xK1fvUlApHGuLc/oSkjLaLb8vKjCPpGntUUpxaQFH/P4+p8Y/vDEf6S/+rpo+HZ21zauDYvF6vSrsaIDtvvSE6M32wVoCwxSSr0OHAaGKqWGikg48G/gPKVUL2AdhmRiCflKqbOUUl8AnwJvKqV6AGcCSSIyHOgA9APOAHqLyGBz344YQ9/dgSzgzsrHr+8JqqqSxCAVn7ZwOIrJTD7MyIdfYNit/2LZR69RcDyHbUt+Jvb0PviHVuzBOIqLyc1IJap9F66c/AbN2nXiz6/fq6+rlVys5omQShkFLR5WQls2Z8/KdUzrfQkJqzZw5UuPIyKMfuVJvnmwqtyz1cNK+7P68sH19zL9rKs44/IL6HiuM4MuTctffV1UpCHa2V3buDYsHp5Ovxoreg7bfVmjlDoIICKbgNYYkojlGYAR0FeaNwtPoPy41pfm/gFAjFLqW4ASyUAzYA8HSibM/DECeCJwQCm10iz/BLgHeKk2h0XkNuA2gKse+g8DL7221hP0CwknJyO19HNuRip+waFVbJq17YTVw4PAiCiCmrUg8+hhju7ZwZH4bWxb8jP2gnyK7UXYvHzod+U4PDy9aNPTuKG17XM2O1f8WqsfznDOnWM569YxAOxfG0dIy+al24JbRHHs8NEK9rlpGRTkHmfTt78AsOHreQyacA1eAf4073YaDyz5AoDAqAju/OE93rr0FjIOHiF+6erSIdKt8xYT26sbuxb90eT9LY++Lhq+nd2pjZ1FLO7fP9UB230pvwS0mOr/lwL8ppQaU0MdueXsqkOA55RSb1coFGkNVX6CnzApgVLqHeAdcC5xSmTr08g8epislCP4hYSxe80yht36SAWb1j0Hsmf1UjoOOp+87Ewyjx4iMCKqgt2ulb+Rsi+e/leNB6BVj/4c3rWZmM5ncGjHJkKiY6kvS9+azdK3ZgPQbcRQhtx9E+u++IE2/XuSn5lN1pGUKvts/nEhpw0ZwK7Fq+g0bBBJ2+PJz8rmoYiyeb8HFn/BnIemkrh+C6l79jP8kYnYfLwpLiyiwzn9WfjK+/8If8ujr4uGb2d3amNnaQqLznTAbnpkAwFAKvAn8KaItFdK7RYRX6CFUuqv8jsopbJE5KCIXKaU+k5EvAAr8AvwrIh8qpTKEZEYoGT5aayIDFRKrQLGUNa7L3/8emGxWjnrujuY9+q/UQ4HHQcNJzSmFWu/m01E6w60PmMALbv25uC2DXz55EQsFgsDRk/A2z+w1nr7XzWeRe+9xB9fvIN3QBBDxt9fX1crsHXeYrqNGMqzu5caj++Mf7h02xMb5zG15wgAvv3X84yf/TKjX51MTkp6BbvqOH4si99ffo/H1v6AUopt8xazdV69lwq4nb/6uqhIQ7Szu7ZxbTSFgK3VutwIEclRSvmLyBDgIaXUJWb5DGCdUmqWiEwC7gKSzHnsc4EXAC+zmn8rpX4QkX1AH6VUqllHBwwpwnCMoDxaKZUgIvcCt5j75gA3YPTo5wHLMOa744GxSqnjlY9f07no1KSa6nC3tJnueF24WxuDa1KTho2Y6vQ9J23eE40yNanuYbsRSil/8+8SYEm58rvLvX8DeKPc50VA32rqal3pczxwbjV2rwGvlS8zh8QdSqnbq7GvcHyNRqNpDDTmx7WcRQdsjUaj0TR5GvPqb2fRAVtTZ5RS+4BuJ7LTaDSaxkJjfr7aWfQctuaUUBepu8bAtgFZp9qFOhPevf2pdqHO7Plx7al2oU60G1lltqnR425tDNBn3sJ6zylHj57h9D0n6eu79Ry2RqPRaDSngqawSlwHbI1Go9E0eXTA1mgakHemjGfE4B6kpGfRc/TkKts7to7i3advpmenVkyeMZdXZv9Suu3uMecx4YrBiAjvz13GG5/9BkBIoB+fvnA7rZqHs/9wKtc9MpNj2a4XNou4+mZ8u/ZEFRZy9OMZFBzYW2G72DyJvvVBbBFRKIeD3C3rSPvuUwACBgwh/IqxFB9LB+DY0gVkrVyIR2g40RMfRsQCVg8yl8wnc3n9M0UppXh56WZW7T2Kl83Kk8N70ykyuIrdzJXbmL/jANkFhSy+69LS8rmb9/JNXAIWEXw8rTw2rCdtwoznceNTMnlh4UZyC+1YRPhgzBC8PFxz42w58S6C+vbHUVDAvpdf5Pie+Co2HZ55DltoGGK1kr1tC4lvvQ4OB83HjiN4wCBwOCjKPMa+l1+kKD2N4AFn0nzseHA4UI5iDrz9Fjnbt7rEX3dsZ3dr49poCovO9By25pTgzBz2Wb1OI+d4Ph8+e0u1ATsiJIDY6DBGDe1FRlZuacDu2i6GT56fyJlj/0NhkZ2f3nyASdM+ZndiMs/dO5r0rFymfziPh8ePICTAl8dfn3NCf+syh+3btSfBQ0dweMZUvNt0IGL0zRx48bEKNmLzxLtNB/L+2gZWD1rcN4X0BXM5vm0jAQOG4N2qHSlfVspWZfVABJTdjnh50+rJlzkw/QmKaxBQcHYO+4+9R/hqUwKvXDaQbUcyeHnJZj4YM6SK3dakdKICfRk969cKgSS3oAg/LxsAy/YkMXdzAq9ePgi7w8FNny3mqQv60CEiiMy8Avy9PLFaap4edHZ+NahPPyIvvZz4yY/h17EzLW+/i533313FzuLjiyPP+EHW7okppC9fRsayxRXKIy+9HO/YViTOeBWLtzeO/HwAfFq3pe1jT7Jt4vga/ajLHHZjaWd3a2NwzRx27LjZTge7xFljG+UctvsnVz0FiEiNSXpNJa2f/k5/XIWIBIvInfWsY5yIND+x5YlZseEvMjJza9yekpHN+u37KLIXVyjv1Caa1VsSyMsvpLjYwfL1uxg11EjrOHJIT2b/aKRAn/3jSi4d2qtKvfXFv0dfsv5cAkD+3ngsvr5YAyv2pFRRoRGsAYrt5Ccm4BEcVnvFxXaU3Q6AeHhUEWM4WZbtSWJE55aICN2iQ8kpLCI1N7+KXbfoUML9vKuUlwQRgPwiOyWZbtfsT6Z9eBAdIgwJxSAfr1qDdV0IHjCItIXG6ELurh14+PljCwmtYlcSMMRqRTxslGTQLSkHsHh7l4pwlASSyuWuwN3a2R3buDbEYnX61VjRQ+IngZO61O5IMHAn8FY96hgHbMVQ7TolbNtziGfuvoLQID/yCoq48KzTWb99HwCRYYEcSc0E4EhqJhGhAS4/vkdwGPZyUoT2jHQ8gsMozjpWrb3Fxxf/7n04uPjn0jL/ngPwad+FwuTDpM6ZVVqfR0gYze98HFtkFKlzZ9fYu64LKbl5RAb4lH6O9PchJSev2qBRE3PiEvh8w26Kih3MuPIsABIzchDg3rkrycgr4PyOLRjb57R6+wtgCw+nMKUsD3dhagq28HCKMtKr2HZ49nn8TutE5vo1ZKxYVloec+PNhA07n+LcXHY9+mBpefDAQcSMuwVbcDDxU55wib/gfu3sjm1cG405EDuL7mGfBCKSY+pPTxeRrSKyRUSuKWcSKCLfish2EfmfiNTYziIyU0TWici2Er1ps7yviPxh6lOvEZEAEbGKyEvm8TabaUARkWEistEs/8DMBV6ieR1uvu8jIkvM90+ZdktEJMHUsQZ4HmgnIpvMc/MXkYWmPvYWERll7t9aRHaIyLum37+KiI+IXAX0AT416yi7O/2N7NybxPRZ85k/8yF+evN+Nv91ALvdcSpcKUcNvQiLhagJ93Ns8TzsqckA5G5Zx75/30Hi1AfJ27mFZjeVDUPaM9JInPog+ybfTeCAc7AGBNXfs+qUFOtYx1U92vLN+OHcdVZXZq3ZCUCxUsQdTuPpi/rwztWDWbr7MGsTk+vtb43U0FOLf/JR4m4YjcVmI7BHz9LyQx9/wOabxpC2ZCGRIy8rLT+2aiXbJo5n97OTiRk7rkHdc7t2buRtXBtNoYetA/bJcwWGRnQP4DxguohEm9v6AQ8CpwPtTNuaeEIp1QfoDpwjIt1FxBND+vJeU5/6PCAPQ5qyDdDT1KH+VES8gVnANUqp0zFGTe5wwv9OwAWmr1NExAY8CuxRSp2hlHoYyAcuN7W0hwL/lTJR3w4Y+tldgWPAlUqpORia29ebdeSVP6CI3Gb+OFnnSN3lhIsnz6zvltP/uqcZNuEFMjJz2Z1oyBgmp2URFW4EuajwIFLSs11yvKBzLiT28enEPj4de2YGHiFlw9seIaHYj1XtlQBEXn87RclJHFtU1rt25OaUDn1nrvgdr9i2VfYrzsyg8PABfNp3Pil/58QlMPaTRYz9ZBHh/t4kZ5f9q5Jz8gj3P7nfWud3bMHSPUmA0YPsGRNOsI8X3jYPzmwTxa7k6kcZnCHiklF0eeNturzxNkXpaXhGlOkte4ZHUJSWVuO+qqiIY3+uInhA1cGx9CULCRl0dpXynK1b8Ipujkdg7YIWteFu7eyObewsFos4/Wqs6IB98pwFfK6UKlZKHQWWUpaze41SKkEpVQx8btrWxNUisgFDc7orhn51RwzxjLVgqGkppewYgft/5nuUUumm7d5yClwfAYOd8P9npVSBKf6RDDSrxkaAaSKyGfgdiClnt1cptcl8vx5Dj7tWlFLvKKX6KKX6WMI7OuHiyRMRYgx1t4wK5bJze/PlgtUA/Lh0I2NHDgJg7MhB/LhkY4111IXMpQtInPYwidMeJiduDYEDhgDg3aYDjrzj1Q6Hh116LVYfX1K+/rBCefn5br/ufSg8cggAj+BQxGasdLX4+uHdrhOFR09u5uGqHm2ZfcO5zL7hXM5p15x5Ow6glGJrUjr+nrY6DdMmZuSUvl+59wgtg/0B6N8qkt2pmeQX2bE7HGw4mFq6qvlkSPnpe7ZPmsj2SRM5tmolYcOGA+DXsTPFublVhmot3t5lc64WC0F9+5F3IBEAr+YxpXbB/c8k7+ABozy6bPmFb7sOiIcNe9bJJ81xt3Z2xzZ2FovV4vSrsaLnsE+e2n6GOaUVLSJtgIeAvkqpDBGZBXibdVe3T3Xltflhp+xHWeU7gzN62tcDEUBvpVSRqfBVUk/l/V0+/D37uYkM7t2R8GB/Eha8xDP/+x6b+ajKu3OW0CwskFWfTibQzweHUky6/nx6XPlvsnPz+fKluwgL9qfIXsw9z39S+ujW9A/n8dkLdzDusrM5kJTGmEdmutptjm/dgF+3XrR6ZgaqsICjH5ctCYh9fDqJ0x7GIziU0IuuojDpILGPvQiUPb4VPHQEft37gqOY4twcjn40AwDPqBaEX3kTxiUgZPz+A4WHE+vt75mtm/HH3iNcNes3vD2s/Ht42UK8sZ8sYvYNhibMG8u38uuuA+QXFTPyvflc2rU1tw7szJy4BNYmJuNhsRDgbWPyBb0BCPT2ZEyv9oz/fAkiMLB1FIPaRNXbX4DMtasJ6tufbu/PxlGQz75Xppdu6/LG22yfNBGLtw/tpzyL2DwRi4WsuI2kzPsRgBbjb8E7piVKKQqTj7J/xqsAhAwaTNiw81F2O47CQhKef9Yl/oL7tbM7tnFtNOaes7Pox7pOAhHJAW4EJgIjgFCMoeD+GEPN8zF6yvvN9+8opb6ppp4ewMdAT4zAuBn4F/AZsBNjmHutiARgDInfgtHLvlYpZReRUOA48Bdwrql5PQvYqJR6TUR+B/6rlJovIq9gDKUPEZGngByl1EumH1uBSzC0rDcopVqZ5fcC7ZVSk0RkKLAIY0ge4CelVDfT7iHAXyn1lIj8CLyslKpVkFenJm14dGrShkenJv17cMVjXZ3u/d7pe87O10Y1yujeePv+jRsFfIsRYOMwAtkjSqkj5vZVGAu4tgJ7TduqlSgVhzEUvg34AFhplhcC1wBviEgc8BtGz/Y9IBHYbJZfp5TKB8YDX4vIFsAB/M88xNPAayKyHKMXXPtJKZUGrDQX0k0HPgX6iMg6jN72TifaZhbwv1O56Eyj0WgqIxZx+tVY0T3sOiIiYZTrhWpODt3Dbnh0D7vh0T3svwdX9LC7PvCj0/ecbS+PbJRRW89h1wEzIcgS4KVT7IpGo9Fo6kBj7jk7iw7YdUApdRg4qYwEIrIa8KpUPFYptaXejmk0Go2mVqwe7j8DrAP234RSqv+p9qEx8dreGrO7NkoSvlh2YqNGRsKpduAkSHjMmScSGw9pr3x+ql2oM+7WxmBkY6ovNaevch90wNZoNBpNk0dclHv/VKIDtkaj0WiaPE3hOWwdsDWNnkEvPEGr8wdjz8tn0Z2PkRq3vYrNpT99jF+zCOymEtBPl08gLzWd7neNo/PYq1DFxeSlprP47ifIOWBkB5uYto307UaCuJyDScwfUy+hslK2r1/N3Pdew1HsYODwSzj/qhsqbF+9cB7fffgWwWFG2sezL76CM4eP5GBCPF/N/C/5x3OxWCwMv/pGep09DIBPXp3K7q1x+Pj5AXD9vY/Tom2Hf6S/JejrouHb2d3auDb0ojONpoGJPX8wwW1b8VmvC2jWpweD/zuFueddU63t77c+TMqmrRXKUjfv4JuhV2HPy6frzdcy8OmH+O3mBwAozsvn67Mvd6m/juJivn77Ze565hWCwyJ46cFb6dZvENGxbSrY9TprGKNvv79CmaeXFzfc/wSRzVuSmZbK9Acm0KlnP3z9jTSro8bfQc9BQ//R/pagr4uGb2d3a+MTYW3EKUedxf3PoAnQGPW1T/a4InKGiIxwlR+tRwxj1xffA3B0XRxeQYH4Nos4wV5lHF6+Gntefun+fs1dkxqzJvbH7yAiOobwqOZ42Gz0OnsYW1avcGrfyJhYIpu3BCAoLBz/oBByapDkdBXu5m8J+rpo+HZ2tzY+EU0hcYruYTcCmpi+9hkYizrnuaIyv+hm5BxKKv2cc/gIftHNOH40pYrt0DenoRzFJPzwK+unV80R3umGq0j8vWy1t9XbiysXz0HZ7Wx49V32/byw3v4eS0shODyy9HNweAT7d+2oYhe3agl7tm0iIqYlV0yYREhERe2V/X9tp9huJzyqTEDh50/e5ZcvZnFaj96MvOl2bKYQyD/J3xL0ddHw7exubXwiGnMgdhYdsBsBZm7yAOBF4CKM1Kf/UUp9aZoEisi3GMpcy4A7lVLVCjyLyEwM1TAfYI5SaopZvg9DyWskYANGK6V2ikg/4FXTPg8Yr5TaVa4+C7ALOFMplWJ+/gsYgCG5OQUj7WkmRp7zZwAfETkLeK7cOZxk21Qtqy4738JbHyI3KRmbvx8XfPw6p107ir/M3gFAh6tHEtmzK99dPLa0bHa3czl+JJmAVi249MePSN/2F1n7DtTH3WolWyqfQ7e+g+g1+DxsNk9WzP+OT16dxqSpr5Vuz0xPZfYr/+H6e5/AYjEGwUbeOJHAkDDs9iK+nDGd37/5lIuuHV8/X93R3xp8BH1duLqd3a6NT0BTWHSmh8QbDw2mr11uW6qpbT0TQyUMjPzgg5VSPYHJwLTylZk/DD7ByCWO6VucKcs5GbjA1Oy+1MyBPhn40tTDrhCsy+thryiseUiv6y3XMXr5t4xe/i25R5Lxj4ku3ebfPIrjR5Kr7JObZJQV5eQSP+cnInuVnXbMOQPp/eDtzB9zJ47CotLyknqy9x/k8Io1hHfvUqNPzhIcHsGx1DL/jqWmEBgaXsHGLzCotBd05vCRHNhTpg2edzyXt595hIuvv5U2nbqWlgeFhiMi2Gye9D9vBIl/Ve2dNXV/9XXR8O3szm18IkTE6VdjRQfsxkND6muXMNf8W16/OghDOGQr8Iq5T2U+wFAnA7gZKBFwXgnMEpFbAeuJTrC8HvZZnsE12m177zO+Pvtyvj77cvb+vJCO144CoFmfHhRkZVcZkhOrFe9Qoz6LhwetLhhC+g5jBWp4986c8+rTzB9zJ3mpZVq+nkGBWDxtAHiHBhPVvycZu3af6BROSGyHTqQcPkjakcPYi4rYsHwhp/ev+O/KTE8tfb9lzUqatTDS0tuLinh/2uP0HXohPc8aWu0+Sik2/7mc6FZt6+2ru/mrr4uGb2d3buMTIRbnX40VPSTeeGhIfe0SSjSsy+tfPwssVkpdLiKtMXKlVzyYUgdE5KiInIshIXq9WX67iPQHLgY2icgZtZzDSZH461JanT+Y6zb+iv14Povverx02+gj2p6/AAAgAElEQVTl3/L12Zdj9fLkkrnvY7F5IBYLB5euYsdHXwMw8JmHsfn5MvwjQ4u35BGSkI7tOOeVp1HKgYiFja++S8auPfX212r14KqJ9/PWUw/icDgYcN7FRMe24edP3yO2fSdO738WS3+cw9Y1K7FYrfgGBHLDfcY5bVyxiN3b4sjNzmLNovlA2WM6H//3WWOhkVLEtGnPNXc+VJsbTdbfEvR10fDt7G5tfCKaQmpSrdbVCGhofW2l1CxzDruPUipVRPoAL5na2N8CnyilvjF1sscppVqLyBDgIaXUJWbdVwJvALOVUv8yy9oppfaY7zdiyHy2wxgev6m2c54Z3MmtLry2q90vNak7ktDfvdJmuuN14W5tDHDHsZ31Hqce+uoyp+85i+8bXOvxRORC4DWMkcX3lFLPV9oei7FmKNi0eVQpVe+FuO7/k6Np0KD62ifgReA5EVlJ7cPaPwD+lA2HgzHPvsUcTl9m+r4Y6GLqYVf/0KZGo9H8zVgs4vSrNkTECryJsUC4CzBGRCpPwv8b+MpcG3Qt8JYrzkEPiZ9iTH3tdGUMdTxsvkpRSi2hmmHqmlBKjauhvHW59+uAIeb7VVRUIHuyhuP2wFhstrNcPdUtfkunbO5do9FoGgUufKyrH7BbKZUAICJfAKOA8mngFBBovg8CDrviwDpgn0LcRV9bRB4F7qBspbhGo9G4FVbXBewYoPwzaAcxpi/L8xTwq4hMAvwwnq6pNzpgn0LcRV/bnJ95/oSGdeD6V652ZXUNzx8zTrUHdeYhv9qe/mucTLrINavf/y5aueF1MdDdvnsuoi4BW0RuA24rV/SOUuqdks3V7FJ5fnwMMEsp9V8RGQjMFpFuNeXPcBYdsN0Ura+t0Wg0zuNZh1XiZnB+p4bNB4GW5T63oOqQ9wTgQrOuVSLiDYQDVR9krwN60ZlGo9FomjxWizj9OgFrgQ4i0kZEPDEWlf1QySYRGAYgIp0xHq+tmtO1jugetkaj0WiaPB4umsNWStlF5G7gF4wnaz5QSm0TkWeAdUqpHzAyU74rIvdjDJePUy54hloHbE2jRinFf3/fyMo9SXjbrEy5uB+dokKr2L21dDM/b91Hdn4Ryx68srR8Q2IyLy/cyO7kTKaOGsiwTsZIVlJmLo/MXUmxUtgdDq7p3YEre7ZvFD4X2ouZ8tNqdh7JIMjHk2mjzqR5sB9FxcVMW7COHUcysAAPnteL3q0iq9RbV7pGBXDNGTFYRFixN40FOyuO2l19RnM6RhhSjp4eQoCXjfu+K1sq4e1h4ekLO7HpUCafbzwEwD1ntyXIx4ZVID41l882HMTVKR+ib5pIwBl9cRQWcHDmy+Tvq5h8Qzy9iL3vMTwjo0E5yFq/mqNfzALAFh5Ji4n3YQ0MojgnmwNvTseengZA1HU3E9CzL4iQs2UjSR+97RJ/3e26cMfvXm24cNEZ5jPV8yqVTS73fjswyGUHNNFD4g1AY5TLrA8i0lpErjsVx/4jIYnEjGzmThzB4xf24flf1ldrd3b75nx00/lVyqMC/ZhycX8u6BJboTzc35v3xw7js5svYNaN5/HRqh2kZOc1Cp+/35xAoLcn395+Mdf17cgbS+IA+HZTAgBfTLiQGdcO4dVFm3DUMwqKwHW9WvD68gSm/LKTvrEhRAdWXMv41abDPPvbLp79bReL4lPZeKhiHvhR3aL5KyW3Qtk7q/bx7K+7eOqXXQR4edCnRc2paE+GgDP64BUVw1/338Khd18nZsLd1dql/jSX+IcmsvvRSfh17IJ/jz4ARF8/gYzlC9n9r7tInvs5UaZghm+Hzvie1oX4R+4i/uE78W17Gn6dT3eJz+50XbjC31Px3asNFw6JnzJ0wG4AmphcJhh5x+sUsM3kAvVmafwhLu7WGhHh9JhwsguKSM2p+uU+PSaccH+fKuXNg/3oEBlcJaG/zWrF08NwsbDYQb2WbrrY52Xxh7n49NYAnNupBWv3H0Upxd60LPq2MuQWQ/288fe2sSMpvcr+daFNqC/JOQWk5hZS7FCsTcygR/OgGu37xYawJjGj9HNsiA+B3h5sP5pdwS7fbrSoVYwbpavT2gX0HkDGckOSMW/3Lqy+fngEh1SwUYUF5G7fbLwvtpO3dw+2sDAAvFrEkrN1EwC52+II7D2gZC8sNhvi4YHYbODhgT3TNdrT7nRduMLfU/Hdqw1PD6vTr8aKDtgNgIjkiMF0EdlqZgMrn/UrUES+FZHtIvI/U7KyprouFJENIhInIgvNslAR+U5ENovInyWKXCLylIh8JCK/isg+EblCRF40j79ARGym3T4ReUFE1piv9mb5LBG5qvx5mG+fB842s5fdLyJW89zWmj5MNO2HiMhiEfkMcMnjZSnZeTQL8C39HBngQ7KLfo0fyTrOmPcXcMmbP3Jj/05EBFS96ZwM9fU5Oft46f4eFgv+XjYy8wrpEBnMsvhD2B0ODh3LYeeRDI5mHa+Xr8E+NtKPl6koHcsrIsTHVq1tqK+NcD9PdiYbl4UAo3vEMCeu+pwQ9w5uy0ujupFf5GD9QdcEvRJsoeEUpZWt4SlKT8VWSf2qPBZfPwJ79SNnq9Erzd+/l6B+hvhGYN8zsfr6YvUP4Hj8TnK2b6bzzE/oPPMTcuLWU3DYNbKP7nRduMLf2mio715tNIUetp7DbjjKy2WGA2tFpCTxcD/KcoMvMG3nVK5ARCKAdzHkL/eKSMkE0tPARqXUZaYgx8fmscDI5T3UrH8VcKVS6hEzZ/jFwHemXZZSqp+I3Iihh31JLefyKBXzit8GZCql+oqIF7BSRH4td27dlFJ7nWum2qmuZ+Yq9buoQF8+n3AhKdl5PDR3BcM6tSTMz/vEO56A+vpcvbILXNq9DftSs7hx1m9EB/rSPSYcq6V+v7mdeaC0hL6xIaw/eKx0LnpI+3C2JmWRkVdUrf1ryxLwsAi3DGhFp0h/dhzNqdbupHBSqxkAi4XYSf8i9ZcfKEo2sv0mffoezcfdQcg555G7YytFaamo4mI8m0XjFdOSnXcZ4nRtHp9KdqduHN+5td4uu9N1UdPxGvt3rzYacyB2Fh2wG45SuUzgqIiUyGVmYcplAohIiVxmlYANDACWlQQ/pVTJONdZwJVm2SIRCRORknHM+UqpIhHZgrGCcYFZvoUySU0wZDpL/r5Sx3MbDnQv1xsPAjoAhea5VRusyycjePXGixk/pFe1lX+1Pp7v4ox5uS7RoRzNLustJGfnEVHN8Ft9iAjwoW14IJsOpJQujKkrrvS5WYAvR7OP0yzQF7vDQU5BEUHenogID5zXs9Tu5tm/0zLU/6T8LSEjr4hQ37IedbCPjWM1BOC+LYP5bMPB0s9tw3zpEO7POe3D8fawYLUI+XYH325JKrWxOxRxhzI5o3lQvQN26PmXEHruBQDkJcRjC4so3WYLDceekVbtfjG33kPBkUOkzf++zK+MdBJfmQqAxcuboH6DcOQdJ3DYReTF78JRkA9Adtw6fDt0OumA7W7XhTt+95zF2oh1rp1FB+yGo95ymWYd1f7QraXOAgCllENEiso9SuCg4v9bVfPejjlNIsbEk2ctfk1SSv1SodBQ+Mqtdg8qJiPI+nByjdOaV/fuwNW9OwCwYvdhvtoQz/DOsWw9nIa/l63a+bK6cjTrOEE+nnjbPMjKL2TzwVSu79vxpOtzpc9nt2/Oz1v20T0mnEU7D9K3VTNEhPwiO0qBj6cHq/cewUMstA2veb7ZGfalHyfS34swP0+O5RXRNzaE9/7cX8WuWYAXvp4eJKSV3cDfX51Y+n5g61Bah/jw7ZYkvDwseHtYyMy3YxHoFh3I7tQaLwunSf/tJ9J/M9ZrBvTsS9jwkWT+sRSf9h0pPp6L/VhGlX2aXX0jVh8/Dr3zWoVya0AgxTnZoBQRo64mfYkxQFSUmmL8KPjeAiL4dT6d1PnfVanXWdztunDH756z6B62pjaWARNF5CMMuczBGMIenYB+YmhX7weuoeaMOquAN0WkTcmQuNnLXoaR1/tZM0imKqWyKi/uOAHXYMxNX2MeB2Af0Bv4CiOZfUnXKxsIKLfvL8AdIrLI7M2fBhyqy8GdZVC7aFYmJHH52z/jbfNg8oh+pduu++AXPrvZ6HG9vjiOX7bvJ7/IzsVv/sCo7m257exubEtK45G5K8nKL2TF7sO8vWIrX91yEfvSsnh10abSX0TX9+9E+0jXrGSur8+jerRlyo9/cvn/fibQx5OpowYCkJ5bwKSvlmIRiPD35emR9U9251Dw+YaD3De4LRYRVu5NJykrn0u7RrE/4zhxh7MAY7HZ2sSqAbE6PK0W7jqrLR4WwSKwMzmHpXtS6+1rebI3riXgjL6c9ur7qIICDr5dNkjU/rk32P3YJDxCw4i8/FryDyXSftrrAKT9+hMZi3/Br/PpRF07DoDcHVs5/OGbAGSuXoFf1+50ePEtUJATt57sDWtc4rM7XReu8PdUfPdqoykEbK2H3QCISDaGUsuLGBJsCviPUupLM8BOxsh6czpG8L2zphyzInIRMA2j55uslDrfnMv+EGgDHAduU0ptFkPPOkcp9ZK5b45Syt98X7pNDG3sDzG0ty3AGKXUbhFpBnxvli3E6EX7m4vVFmDMxc/C0IH9DzASo7edAlyGocNdOtddG7X1sDWuwS1ziX/7+Kl2oU60Gt7nVLvwjyBw/DP1jraP/bzd6XvOcxd3aZTRXfewXUwDyGXOB+ZXKkvH6AFXtn2q0mf/mrYBbyqlnq5kfxRj3ryEx8zyIsw0e+V43HyVZwl1ODeNRqP5u2gKPWwdsF2Iu8hlajQazT8NHbA1FXAjuczWrq5To9FoGjM6YGtcxj9NLvOvr1eeahfqRLuRfU+1C3Xm4TX/OdUu1JmuO6NOtQt14ujgwlPtQp3Z8+PaU+1Cnekzvv516ICt0Wg0Go0boAO2RqPRaDRugKeH+2fi1gFb0+hpOfEugvr2x1FQwL6XX+T4nvgqNh2eeQ5baBhitZK9bQuJb70ODgfNx44jeMAgcDgoyjzGvpdfpCg9jeABZ9J87HhwOFCOYg68/RY52+uffhKMFJkvL93Mqr1H8bJZeXJ4bzpV85zpzJXbmL/jANkFhSy+69LS8rmb9/JNXAIWEXw8rTw2rCdtwgIBiE/J5IWFG8kttGMR4YMxQ/BygVhBxNU349u1J6qwkKMfz6DgQMVkdWLzJPrWB7FFRKEcDnK3rCPtu08BCBgwhPArxlJ8zEjEd2zpArJWLsQjNJzoiQ8jYgGrB5lL5pO5/Ncqx64r70wZz4jBPUhJz6Ln6MlVtndsHcW7T99Mz06tmDxjLq/MLsvvc/eY85hwxWBEhPfnLuONz34DICTQj09fuJ1WzcPZfziV6x6ZybHs+ufjLo87Xhfu9t2rjabQwz7hc9hmUoyZQDOlVDdTaOJSpZT7TZBpGg3rRgxz6pnIoD79iLz0cuInP4Zfx860vP0udt5fVUrR4uOLI8+4wbZ7Ygrpy5eRsWxxhfLISy/HO7YViTNexeLtjSPfSD/p07otbR97km0Ta54oq8sc9h97j/DVpgReuWwg245k8PKSzXwwZkgVu61J6UQF+jJ61q8Vbsy5BUX4eRk5a5btSWLu5gRevXwQdoeDmz5bzFMX9KFDRBCZeQX4e3nWeCNK3bzbKX99u/YkeOgIDs+YinebDkSMvpkDLz5WwUZsnni36UDeX9vA6kGL+6aQvmAux7dtJGDAELxbtSPly/crVmz1QASU3Y54edPqyZc5MP0JijNrTsDS9c/AE/p7Vq/TyDmez4fP3lJtwI4ICSA2OoxRQ3uRkZVbGrC7tovhk+cncubY/1BYZOenNx9g0rSP2Z2YzHP3jiY9K5fpH87j4fEjCAnw5fHXq8sWXJGjtzmvO91Yrgtn57Aby3cPoM+8hfWOtjNW7XX6Oey7B7ZplNHdmTGCdzGexy0CUEptBq5tSKfclYbSwRaRcSIy4+Q9+3vqbAiCBwwibaHRK8vdtQMPP39sIaFV7EpuDGK1Ih42SrKtlpQDWLy9KVGuKLlhVC53Bcv2JDGic0tEhG7RoeQUFpGam1/Frlt0KOHVCB6U3JQB8ovslGSiXbM/mfbhQXSIMNJOBvl4uaTX4N+jL1l/LjGOtzcei68v1sCKPT9VVGgEa4BiO/mJCXgEh9VecbEdZbcDIB4eLlOOWLHhLzIya051mpKRzfrt+yiyF1co79QmmtVbEsjLL6S42MHy9bsYNdTIZz9ySE9m/2gshJz940ouHVp9nvv64G7XhTt+92rDKuL0q7HizJC4r1JqTaW0l/YG8setaYI62CeNiFhN4ZN6YQsPpzClTEaxMDUFW3g4RRlV9X47PPs8fqd1InP9GjJWLCstj7nxZsKGnU9xbi67Hn2wtDx44CBixt2CLTiY+ClP1NfVUlJy84gsJxcY6e9DSk5etTfhmpgTl8DnG3ZTVOxgxpWGDGRiRg4C3Dt3JRl5BZzfsQVj+5zUU4QV8AgOqyCcYc9IxyM4jOKs6iUxLT6++Hfvw8HFP5eW+fccgE/7LhQmHyZ1zqzS+jxCwmh+5+PYIqNInTu71t51Q7NtzyGeufsKQoP8yCso4sKzTmf99n0ARIYFciQ1E4AjqZlEhAbUUtPJ4W7XhTt+92rD0ogDsbM408NOFZF2mD+bTIWmpNp3+WfiYh3s8SLyl6nyNahceYSIfGNqUa8VkUEiYjE1roPL2e0WkWbV2VdzrFYistDUtl4oIrFm+SzTz+WmLyXymn+bHna11PCLPP7JR4m7YTQWm43AHmXqRYc+/oDNN40hbclCIkdeVlp+bNVKtk0cz+5nJxMzdlyDulfXW8VVPdryzfjh3HVWV2at2QlAsVLEHU7j6Yv68M7Vg1m6+zBrE5Pr73C11CxVGTXhfo4tnoc91Th27pZ17Pv3HSROfZC8nVtodlPZsKk9I43EqQ+yb/LdBA44B2tA/cRK6sPOvUlMnzWf+TMf4qc372fzXwew26vNCNwgNInropF/92rDZhWnX40VZwL2XcDbQCcROQTcB9zRoF65N+V1sM8DpotItLmtH/AgRg7xdqZtFUz7pzEC9fkY2tYlvAa8opTqiyGx+Z6Zh/x74HJz//7APjPVaBX7ag45A/hYKdUd+BR4vdy21sA5GFra/xMRb2ACph42hmToraaYSck5PqGUKu9zyXndJiLrRGTd3MSatUIiLhlFlzfepssbb1OUnoZnRJmMomd4BEVp1csoAqiiIo79uYrgAVUHO9KXLCRk0NlVynO2bsErujkegSeeP62JOXEJjP1kEWM/WUS4vzfJ2Xml25Jz8k5a5ej8ji1Yusf4fRzp70PPmHCCfbzwtnlwZpsodiVX3ws+EUHnXEjs49OJfXw69swMPELKhrc9QkKxH6vaiwKIvP52ipKTOLaorHftyM0pHfrOXPE7XrFtq+xXnJlB4eED+LTvfFL+uopZ3y2n/3VPM2zCC2Rk5rI78SgAyWlZRJkKV1HhQaSkZ7vkeO52Xbjjd89ZmsKQ+AkDtlIqQSl1HhABdFJKnaWU2tfgnrkvpTrYZsAs0cEGUwfbHCou0cGujv7AEqVUilKqEPiy3LbzgBkisgn4AaPXHmDalPTmry23T0325RkIfGa+n13Jr6+UUg6lVDyQgKE2Nhy40axzNRCGoYddco7V6mErpd5RSvVRSvW5IjamhlOHlJ++Z/ukiWyfNJFjq1YSNmw4AH4dO1Ocm1tlSM7i7V02t2axENS3H3kHDOlHr+ZlxwnufyZ5Bw8Y5dHNS8t923VAPGzYs7Jq9OlEXNWjLbNvOJfZN5zLOe2aM2/HAZRSbE1Kx9/TVqdhz8SMMt3olXuP0DLYSAnfv1Uku1MzyS+yY3c42HAwtXSVcF3JXLqAxGkPkzjtYXLi1hA4YAgA3m064Mg7Xu1weNil12L18SXl6w8rlJef7/br3ofCI8aPMY/gUMRmKLRafP3wbteJwqOHT8pfVxERYlz6LaNCuezc3ny5YDUAPy7dyNiRxuDT2JGD+HHJRpccz92uC3f87jmLRcTpV2OlxjlsEXmghnIAlFIvN5BP7o4rdLBr22YBBiql8soXisgqoL2IRGAoZ/3nBPa1HLparezyn09KD/tkyFy7mqC+/en2/mwcBfnse2V66bYub7zN9kkTsXj70H7Ks4jNE7FYyIrbSMq8HwFoMf4WvGNaopSiMPko+2e8CkDIoMGEDTsfZbfjKCwk4flnXebzma2b8cfeI1w16ze8Paz8e3jZAqaxnyxi9g3nAvDG8q38uusA+UXFjHxvPpd2bc2tAzszJy6BtYnJeFgsBHjbmHxBbwACvT0Z06s94z9fgggMbB3FoDb1zwx2fOsG/Lr1otUzM1CFBRz9+K3SbbGPTydx2sN4BIcSetFVFCYdJPaxF4Gyx7eCh47Ar3tfcBRTnJvD0Y+MtYyeUS0Iv/ImSi6ZjN9/oPBwYjUe1I3Zz01kcO+OhAf7k7DgJZ753/fYzEeY3p2zhGZhgaz6dDKBfj44lGLS9efT48p/k52bz5cv3UVYsD9F9mLuef6T0ke3pn84j89euINxl53NgaQ0xjwys95+Vsbdrgt3/O7VRiMe6XaaGh/rEpEpte1YWelJY8xhAzcCEzGkK0OBdRg95k4YqltdMHSw5wPvKKW+qaaeaOBPoBeQBSwC4pRSd5vzwxuVUtNN2zOUUpvM99OBKCBMKTXCLKvWXkTGAX3MOn8AvlZKzTbLRymlLheRWUAkcAmGlOdSoL15jiOA0ZX0sPvipLyms491NRbcMTWps491NSaceayrMVGXx7oaC26ZmtQFj3V9EXfI6XvOtT1iGmV4r7GHrQPySaGAbzGGmOPMz48opY6ISCdgFfA8ZTrY31ZbiVJJpn71KowFfhuAkiwI9wBvishmjP/fMuB2c9uXwFpgXLnqarMvb/OBiDyMoW1d/qHIXRiBuhlwu1IqX0Tew5jb3iBGV71ED1uj0WgaJY15qNtZTvhYl4i0xVi4NAAjAK0C7ldKJTSwb25FA+hgfwh8WE15KmVz1ZW3raPSkHxN9kqpWcAs8/0+4NwaXFmplLq/0r4OtB62RqNxI2xNINOZM6vEPwO+AqKB5sDXGAumNCamDvYqtA62RqPRNEqsFnH61VhxJnGKKKVml/v8iYhUzU/3D8ZddLDrilJqXEPV7RVU+ZQbN/8Kc7/kfi/1m3uqXagzqf1OtQdNn9NGV0nF8I+gSQ+Ji0hJDrrFIvIo8AXGkPg1wM817aepG/80HWyNRqM5FTSFVeK19bDXU/YIDxgrn0tQwN+zFl+j0Wg0mnrSpHvYSqk2NW3TaDQajcadaMwpR53FKT1sEemG8fxwaVoepdTHDeWURlOe6JsmEnBGXxyFBRyc+TL5+/ZU2C6eXsTe9xiekdGgHGStX83RL2YBYAuPpMXE+7AGBlGck82BN6djTzfSK0ZddzMBPfuCCDlbNpL00dsu8bdrVADXnBGDRYQVe9NYsLNiXuerz2hOxwgj45anhxDgZeO+78qWLHh7WHj6wk5sOpTJ5xuNrGH3nN2WIB8bVoH41Fw+23DQZSJHSin++/tGVu5JwttmZcrF/egUVVWV6a2lm/l56z6y84tY9uCVpeWF9mKm/LSanUcyCPLxZNqoM2ke7EdRcTHTFqxjx5EMLMCD5/WidyvXPLdcX583JCbz8sKN7E7OZOqogQzr1BKApMxcHpm7kmKlsDscXNO7A1f2bN8ofP6729kd27g2GnPKUWc54SpxM4HKG+ZrKPAicGmtO2k0LiLgjD54RcXw1/23cOjd14mZUP16x9Sf5hL/0ER2PzoJv45d8O/RB4Do6yeQsXwhu/91F8lzPyfqWuMRc98OnfE9rQvxj9xF/MN34tv2NPw6n15vf0Xgul4teH15AlN+2Unf2BCiAysusPtq02Ge/W0Xz/62i0XxqWw8VDEN6Khu0fyVUjFh3Dur9vHsr7t46pddBHh50KdFRfnL+vBHQhKJGdnMnTiCxy/sw/O/rK/W7uz2zfnopvOrlH+/OYFAb0++vf1iruvbkTeWxAHw7Sbjyc8vJlzIjGuH8OqiTThc9Cujvj5HBfox5eL+XNAltkJ5uL83748dxmc3X8CsG8/jo1U7SMnOq7L/qfD5725nd2zj2mgKqUmdeazrKmAYcEQpNR5D1MK9lvg2AlyplS0io0Vkh6mM1UdEXj/xXq6htvMoZ5NzIhtnCeg9gIzlCwHI270Lq68fHsEhFWxUYQG52zcb74vt5O3dgy3MELPwahFLztZNAORuiyOw94CSvbDYbIiHB2KzgYcH9syTE0woT5tQX5JzCkjNLaTYoVibmEGP5jUrVPWLDWFNYpnkZGyID4HeHmw/WlF8It9UlbKK8XiKK9PELY0/xMXdWiMinB4TTnZBEak5VW+gp8eEVytWsSz+MBef3hqAczu1YO3+oyil2JuWRd9WzQAI9fPG39vGjqTqRUX+bp+bB/vRITK4Sopem9WKp5nmtLDYgSu1vNytnd2xjWvDanH+1VhxxrU8M1GGXUQCgWSgqhyPplZcrJU9AbhTKTVUKbVOKXWPC+uulb9b89sWGk5RWpkmb1F6KrbQ8BrtLb5+BPbqR85Wo/eRv38vQf0MLZPAvmdi9fXF6h/A8fid5GzfTOeZn9B55ifkxK2n4PCBevsb7GMj/XhR6edjeUWE+NiqtQ31tRHu58nOZOP3jQCje8QwJ656gYx7B7flpVHdyC9ysP5g/X9clJCSnUezAN/Sz5EBPhVUpU5Ecvbx0v09LBb8vWxk5hXSITKYZfGHsDscHDqWw84jGRzNOt4ofK6NI1nHGfP+Ai5580du7N+JiICTU9SqjLu1szu2cW38U3rY60yd5XcxVo5vANY0qFdNEFdpZYvIZAw1rf+ZdZX2zkXkKRH5QESWiEiCiNxTbr/vRGFA7SYAACAASURBVGS9iGwTkdsq+TVVROJE5E8RaWaWNzP9iTNfZ5bYm3/9Te3sDea5jHJ5o0G1Uio15b/HYiF20r9I/eUHipKPAJD06Xv4de5G++fewK/z6RSlpaKKi/FsFo1XTEt23nUjO+8ci3/XHvh26tYQ7tbYG+4bG8L6g8dK56KHtA9na1IWGXlF1dq/tiyBh3/Yhs0qdIr0r7evtflXl3tWtecncGn3NkQG+HLjrN94+feNdI8Jx2pxTfelvj7XRlSgL59PuJBvJ17Mz1v3kZab75J63a2d3bGNa6MpyGuecNGZUupO8+3/RGQBEKiU2tywbjVZymtlhwNrRWSZua0fZcIgC0zbOZUrUEo9IyLnYohsrDMVssrTCWOtQQCwS0RmKqWKgJuVUuki4mMe9xulVBrgB/yplHpCRF4EbsVQ+nodWGqKgFiByhEiH7hcKZUlIuHAnyLyg6oxmhp62MBtAJP7dOWq9rHV2oWefwmh514AQF5CPLawMk1eW2g49ozqNXljbr2HgiOHSJv/fWmZPSOdxFemAmDx8iao3yAcecf/z955h0dRrQ/4/dI7ISShd5COIKAgwsV6bYhcewfFi/fau169iHqx/+y9oij2gooKFrqCSC+CICCdEAgJgfT9fn/MbLJJdjcbskl2wnmfZ57szp6Z+XYyu2fPmXO+l6QTTyNv3VpcBdYXxf5lvxHXuSsH16z0FX5AZOUVkRJX1qJOjo1kn48KeEDrZCYv3lr6vEOTODqnJvC3TqnERIQRHibkF7v4fMWOsvfjUpZty6ZPi0b8vuvQ7zx8tGgdXyyz7n12b57Crv1lLbKM/XmkVcPT3DQxjl37D9I0KY5il4vcgiIaxUQhItxyUt/ScldO+oHWKYf+QyOYMQdCWmIsHVKTWLpld+mAqeritPPsxHMcKA16lLiIdFXVNSJylJfXjlLVxbUbWoOk1JUN7BIRtys7B9uVDSAibld2pQo7AKaqagFQICIZWNKOrcANIjLSLtMay1+9BygE3PfPFwHu0SMnYFm5sOPNrnAcAR4SkaGAC2hpH2unr8BU9VXgVYAVF53us2Lf+/3X7P3eCimx7wCanDKc7J9nEdupCyUHD1C8L6vSNk3Pv5zw2Hi2vfpMufXhiUmU5O4HVdJGnM/emdMBKMrcbf0omBIGIsR360Xmt1/4CilgNu09SHpCNE3io9iXV8SANo15ff5fleNNjCYuKoINe8q+EN9YUKaeHNQuhXaNY/l8xQ6iI8KIiQgjO7+YMIGezZNYn1kzi+n5/Tpzfj9LYT53/XY+WryOU7q1YeX2PSRER3q9J+mLIZ1aMHXFJnq3TOWnNVsZ0LYpIkJ+UTGqEBsVwYKNO4mQMDqk+r6fX5cx+2JXzkEaxUYRExlBTn4hy7dmcsmALiERc12cZyee40AJ5a7uQPHXwr4FqzX0f15eU3zLIgy+CZYr2x8FHo9LgAi7FX4Slhf7oIjMpGyKXpFHq7iEAKf6AZcAaUA/W7G5yWOfQWP/koUk9hnAEU+/gRYUsPWVp0pf6/Twc6y/+3oiUpqQPvJC8rdtptND1vi7PdO/JmvGNOK79aLZhaMAOPD7Sra/9QIA2QvmEt+jN50fexEUcpctYv/imt/pcSm8v3grNw3tQJgI8zbuZUdOPmf1aMZfWQdZtj0HsAabLdxc+YeHN6LCw7j2uA5EhAlhAmsycpn1Z2aNY3UzuGNz5m3YwchXphITGcG408vyg1785jQmX2n1djw7YxnTVv9FflExZ7zwJSN6d+CfQ3oy4sgO3PfVfEa+PJWk2CgmjBgEwN4DBVz/0SzCBNIS4rh/ePCS+tU05lU79nDHZ/PIyS9k7vrtvDJ3JR+NOY1Ne3J4+qelCNYH8JJjutIpPTgj8p12np14jv3RAOpr3z5sAPs+6iBVnVd3ITVMguXKtvc1k/Jd4rep6pm2kjNXVZ+wy63EclkfCYxR1eG25nMpcKqqzhSRXFVNsMufC5ypqqNE5AOsrvKn7S7xeLv7O1dVE0TkRqCTql4vIsdjObvbq+omz336wl8LOxR5buRD9R1CtXnigPNyiRsM3kga/UCNq9tVO3IC/s7p0TwpJKt3vyMT7NHhxkAVHNyu7OVYruyfsF3Z9utuV/ZKYCM+XNmHyHdYLe3lWCll5wewzY3A8SKyAqurvEeF198D+ovIb1it7TVBjNdgMBiCikjgS9X7klNFZK2IrLddG77KnSsiKiL9g/EeAun+nC4i5wCf+RtQZPBNLbiyh3nbVlXHVyjnOez5NB/7SvB4/An2fXNV3QVUGvntLm97tgdVtU+DwWAIBYJlzbR7HF/AGu+zFWsQ75equrpCuUTgBmBBcI4c2LSuW7Ac2AUikiMi+0UkJ1gBNHSMK9tgMBjqnyBO6zoaWK+qG1S1EMtk6W1a64NYmUGDNmctkGldicE62OFIQ3Vl15S2pwSlh6jOMPeDDd64Lf4f9R1Ctbn+8//UdwjVptfomu8jiIPOWgKeWZa2Yo1F8jiW9AVaq+rXInJbsA4cqPyjMdY0IE/5x2zfWxiCgXFlGwwGQ3CoTpe4Z84Im1ftaalQRX4ke7D2U8CoagdZBVVW2CIyBmsAUius0cUDsbp4zbQug8FgMDiC6jSwPXNGeGErVi4LN60Az3zCiUBPYKadR70Z8KWInKWqv1UjjEoE0sK+ESu5x3xVPd6eFnR/TQ5qMASKExV/h5tG0eg1A8Np2lVwntrWH0FMnLIQ6Cwi7YFtwIXAxe4XVTUbK5MlUH4abk0PHMigs3xVzbcPHK2qa4DaT0tjMOBMxd/hplE0es2qcaJ21Wlq26oIl8AXf6hqMXAdMA34HfhIVVeJyAMiUqvq6UAq7K22/OML4HsRmUL55r+hjpAgKjqdghMVf4ebRtHoNavGidpVp6ltq0JEAl6qQlW/UdUjVLWjqk6w141T1S+9lB0WjNY1BFBhq+pIVd1nz/H9L/AGcHYwDm6oHnWttgwFnKj4O9w0ikavWTVO1K46TW1bFWES+BKq+KywRWSqiFwiIvHudao6S1W/tOeeGeoYCZKi02Nf3rSaaSLyqYgstJfB9voVIpJsH3+PiFxur58kIieJSA8R+VVElorIchHpHIz37ETF3+GmUTR6zapxonbVaWrbqghmprP6wt+gs1exbqY/LSI/Ae8D35jKut6psaLTxpdW8xngKVWdKyJtsO7TdAPmAYPtfW8AhgDvYM0a+BfwMPCMqr4nIlFAeMUDek6VePryMxg9rJIIDnCm4u9w1igavWbVOEW76mS1bVUE56di/eKzwlbVKcAUsfzJZwFXYDmxv8FSRH5fRzEayhMsRacvreZJQHeP+zhJdoq9OcBQrAr7JeCfItISK+Vqroj8AtwjIq2w0tiuq3hAz6kSOW+N83m7zYmKv8NZo2j0mlXjFO2qk9W2VRHIvelQJ5BMZ3nAh8CHItIbeBur8q7UgjLUCcFSdPrSaroNbeVuCNqt+GuBNsA9wEjgXKyKHFWdbGdmOwOYJiJjVPWnAN6PX5yo+DvcNIpGr1k1TtSuOk1tWxVVjf52An71mgD2vc3zsbrHm2PlFX9fVZfWfngGTyS4ik5fWs3JwBJVfdx+rY/7fy0ifwA5qtpfRO7EmtpwnapOEZEOwEZVVRF5Gtikqk/7ei/+WtgGg1MwqUnrhl7vf1Pj6nZf7sGAv3OSE+JCsnr3Nyjpavve9WKsXNh3qGoHVb3TVNb1Rl0oOm/A0mYuF5HVwDUery0A/rAfz8HKqTvXfn4BsFJElmL9eHjnEI5tMBgMtUIwp3XVF/66xI/F+vL/wfZiG+oRCb6i05dWMxOr8vW2zWUej3/G4wefqj6MNfDMYDAYQo5Qnq4VKP4GnQXBj2IIBmIpOmdiFJ0Gg8FwSDSA+jowW5ehfmmIis4/Pp5Xn4evNh2HD6jvEA4L/vxqYX2HUC1ub7u6vkOoNj3WNKvvEKpNMOYSBzGXeL1hKuwGjlF0GgwGA4QFNXFr/eB3LrmIhIlI7c5mNxgMBoOhtlFX4EuI4reFraouO3VlG1Xd7K+swWAwGAyhioRwRRwogXSJNwdWicivQGkaHVWtVY2YweCm9dhraTTgGFwFBWx68jEO/lkpiRqdH3iYyJQmSHg4+1etYPOLz4LLRYvLRpE8cDC4XBRl72PTk49RtHcPyQOPpcVlo8HlQl0lbHnlRXJXB6czSVV5ctZyftm4i+jIcP57Sj+6ekm+8dK8VXz7+xb2FxQy49qyj9Nnyzfy6bINhIkQGxXO3Sf2pX2TJADW7c7m0R+XcKCwmDAR3rxoGNERNcth5LR43TjtugBIO/9K4nr0RQsL2fXO8xRs2VjudYmMovnVtxKZ1gx1uTiw4jf2fPEeAIkDh5H6j8so2WcZz/bN+o6ceT8SkZJK87G3IxIG4RFkz/yW7DnTaxzrq/eN5vShR7J7bw59zxtX6fUu7Zrx2v1X0rdrW8Y9/xlPTZpW+tp1F53EVf8YiojwxmezeW6ylRizcVI87z16DW1bpPLX9kwuvuMl9u0PjhCmShpAhR1I4pS/eVuvqrNqJSLDYcFvp58Y0A2lRv2PJv2skawbdzfxXbrR+pprWXNzZS9vWGwcrjzrg9/xnvvYO2c2WbNnlFufftZIYtq0ZfPzTxMWE4Mr38plHNuuAx3u/i+rxvqeGFGdQWc/b9zJR0s38NTZg1i1M4snZy7nzYuGVSq3csdemiXFcd7E6eUqwAMFRcRHW3mnZ/+5g8+Wb+DpkYMpdrm4YvIMxv+9P53TGpGdV0BCdBThNZyvEkrxBjroLFSui0ZtG/t8rSJxPfqSfPzpbH9+AjHtO5N23pVseezucmUkMoqY9p3J+2MVhEfQ6qb72PvdZxxctYTEgcOIaduR3R++UX7H4RGIgBYXI9ExtP3vk2x5/B5Ksr1nTOsxPymgeI876ghyD+bz1oNjvFbYaY0TadO8CSOOP4qsnAOlFXaPji1595GxHHvZ/ygsKubrF27h+ofeYf3mDB6+8Tz25hzg8be+4fbRp9M4MY7/POsre3IZhUverPGIsYL9+wK+iR2dmBySI9QC0WvOAtYAifbyu6msDx0nOq1FpI+InF4fx04eOJg9P1qthQNrfyciPoHIximVyrm/fCU8HImIxJ2V1b0eICwmBrcCyf2lXHF9MJj95w5O79YaEaFn8xRyC4vI9GJ86tk8hdT4mErr3ZUfQH5RMe4JKb/+lUGn1EZ0TrPycTeKja5xZe3EeMGZ10XCkQPImT8TgPyN6wiLiyM8qXxPhhYVWpU1QEkx+Zs3EJHcxP+OS4rR4mIAJCIiaLqpuYv/ICvbd27y3Vn7WbR6E0XFJeXWd23fnAUrNpCXX0hJiYs5i9Yy4nhL9DN8WF8mfWXNEJn01TzOOt67AKg2EFdxwEuoUmWXuIicDzyONQ9YgOdE5HY72YahmtSW01pEwm0hSLD3G4FlB+sPfFOd7VS1xld+ZGoqhbvLnLyFmbuJTE2lKGtvpbKdH3yE+CO6kr3oV7Lmzi5d3/LyK2ly4smUHDjA2rtuLV2fPGgwLUeNITI5mXX33VPTUEvZfSCPdA+HcnpCLLtz87xWdr74ZNkG3l+8nqISF8+fYzmFN2flIsCNn80jK6+Ak7u04rL+hzTbz9HxgjOvi4jkJuVsV8VZe4lIbkJJjneHdVhsHAm9+7N1xtTSdQl9BxLbqTuFGdvJ/GRi6f4iGjehxb//Q2R6MzI/m+SzdV0XrPpzGw9c9w9SGsWTV1DEqcf1YtHqTQCkN0liZ2Y2ADszs0lLSay7wBpAl3ggxrF7gAGqeoWqXo6lcPxv7YbVcKkFp/UD9lzrQSLST0RmicgiEZkmIs3tcjNF5GkR+dk+5tH2+hQR+cJOQzrflrsgIuNF5FURmY6VYvQB4ALbdX2BiMSLyJu2L3uJiIywtxslIh+LyFdAzW+i+cJHq2fdf+9i2aXnERYZSdKRZVrHbe+8yfIrLmLPzB9JH3526fp9v8xj1djRrH9wHC0vG1Wr4VW3zXPukR34dPQpXHtcDyb+ugaAElWWbd/D/af159XzhzJr/XYWbs447OL1SYhfFz6C9r46LIxmV93MvhnfUJxpnbMDK35j073/YvOEW8lbs4KmV5TdAijO2sPmCbeyadx1JA38G+GJh25FqylrNu7g8Ynf8u1Lt/H1Czez/I8tFBeHQGXZ0EeJ24SpquenbA8NQy1anwTTab1SVceJSCQwCxihqrvtHwETgCvdZVX1WBEZCrwJ9ATuxxJ9nC0iJ2BVzn3s8v2A41Q1T0RGAf1V9ToAEXkI+ElVrxSRZOBXEfnB3m4Q0FtVKzV1PH3Yd/fowj/atPT6ptLOHEHa360e+APr1hKVVubkjUpNo2iPdycvgBYVsW/+LyQPPJacJYvKvbZ35o90Hv8Q2997u9z63JUriG7egoikJIpzcnzu2x+fLNvAlBWbAOjWLJmM/WWys4zcvENWP57cpRWP/WSl7k9PiKVvy1SSY608OMe2b8bajH0MaJPe4OMFZ14Xjf52Ko0GnwhA/l9/EtG4rHs7onEKxfsq9wgApF9yDUUZO9j3U1nr2nWgzHOdPfcHmoy8pNJ2JdlZFG7fQmynbuQumX9IMQeDiV/MYeIXcwB48Lp/sHWX1eLP2JNDs9RG7MzMpllqI3bv3V93QYVwRRwogVS839mttVH2F/dULBOU4dApdVqr6i6sitY9qulXVd1gd2+7nda+KAHcNq4uWJXw97aA416glUfZ9wFUdTZWKz7Z3vcke/1PQBMRcf80/7KiYtODU4C77OPMBGKwtJsA33urrO1jvKqq/VW1v6/KGmD311NYff1YVl8/ln2/zKPJiacAEN+lGyUHDlTq9gyLiSm7fxkWRqMBR5O3xZqFGN2i7DjJxxxL3tYt1vrmLUrXx3XsjEREHvKXMlgtzEmXnsCkS0/gbx1b8M3vW1BVVu7YS0JUZLW6lzdnlX0xz9u4k9bJVtr3Y9qmsz4zm/yiYopdLhZvzSwdjd3Q4wVnXhfZs75j80O3s/mh28ld9itJA4cBENO+M668g167w5ucdSHhsXHs/vitcus973fH9+5P4c5tAEQkpyCRUdbbjIsnpmNXCndtP+SYg0FaY6uru3WzFM4+oR8ffrcAgK9mLeGy4YMBuGz4YL6auaTugnK5Al9ClEB82LeLyD+wvtwFS9l4KBYoQxnBclrne9y3FmCVqg6qxn69xeEu53u0ibXdOaq6ttxKkWOq2K7aZC9cQKMBx9DzjUm4CvLZ9NTjpa91f+4VVl8/lrCYWDrd9yASGYWEhZGzbAm7v/kKgFajxxDTsjWqSmHGLv563jJ+Nh48lCYnnowWF+MqLGTDIw8GLeZj2zXl5407OXfi98REhHPvKWUDay579ycmXXoCAM/NWcn0tVvILyph+OvfclaPdlw9qBufLNvAws0ZRISFkRgTybi/9wMgKSaKi47qxOj3ZyICg9o1Y3D7mqeZdFq84Mzr4uDKxcT3PIq2DzyPFhaw650XS19r85/H2fzQ7UQkp5By2rkU7thKm7sfA8qmbyUffzrxvQeAq4SSA7nsevt5AKKatSL1nCtwf6SzfviSwu01T5sx6eGxDO3XhdTkBDZ89wQPvDyFSHtK3mufzKRpkyR+eW8cSfGxuFS5/pKTOfKce9l/IJ8Pn7iWJskJFBWXcMMj75ZO3Xr8rW+Y/Oi/GHX2ELbs2MNFd7xU4zgDpSHMww5kWtejqnpnVesMgVGLTusoYDVWnvBf7C7yI1R1lYjMBNao6jUichzwkqr2EpFngd2q+qCIDAOeUtW+IjIeyFXVJ+x9nwOcpapX2M8fApKA623/dV9VXVKx69wfgU7rChVMLvG6wWm5xKszrStUCHRaVygRjGldRbs2BvydE9m0vTOndQEne1l3WrADOYyoFae1qhYC5wKPisgyYCmWItVNlj2l7GXgKnvdeGz3tX3MK3zsfgbQ3T3oDHgQiASW26lrg9cMMRgMhtqgIQ86E5F/Af8GOtpf6G4SAZ9ziQ2+qU2ntf18KTDUR/FPVfXuCuX3AiO87He8l3IVm5hjvWw3EZhYRdgGg8FQ5zSELnF/97AnY3XJPgzc5bF+v69BRQbfGKe1wWAw1CMNucJW1WwgW0SewWoV7gcQkUQROUZVF9RVkA2B+nRaq+qwQzlubXLtwLuqLhRCPHVa3WVkChbpT/yrvkOoNrtW7K66UAiR3KGKLGQhyOeZv9V3CPVDQ66wPXgJ8Py2OuBlnaEWMU5rg8FgqCGHSYUt6jGU3FZuBrKdwWAwGAwhgZSEbo7wQAmk4t0gIjdgtarBGoi2ofZCMhjKGNixCbf8vQthYcKXS7bxzrxN5V5vmhTDfWf3ICE6grAw4cUf1/Pz+kwiwoS7z+xG1+ZJqMKT09ay+K8soiPCePi83rRsHIfLpcxZt5sXf1wf1JhXLJzP5JefxlVSwtDThnPGBZd7Lbdwzk+8+L97GffcG7Q/ohvFRUW8/cyjbFy3hjAJ4+J/3UTXI62OrE/fepl5P3zHwdz9vDzlx6DGC5B67mhb+1hAxqQXKdhaWfvY7KpbiExtiqqLgysWsefLyeXKxPc5huZjbmXLY3dRsHkDESlptLn3KYoyrCQe+ZvWsfuD14IWc/f77yT9hCGU5OWz7Jb/krPyd59l+7/5LHFtWjH7pH+UW99h7BV0u/dWpvceSlGWlcQkZWB/uo+/g7CICAqz9jH/vCu97fKQcNp5duI59kkQRS71RSAV9jXAs1iZsxT4ETu9pMFQm4QJ3H5aV65/dzEZOflMHHMMc9buZmNmWW6WK4e054dVu/hs0Vbap8bz5MV9GfnsXM4+yspkdckr82kcF8nTFx/FqNetYRfv/fIXizZlEREmvHB5PwZ1asIv632ntawOrpISJr3wBLc9/Awpqek8cP1V9Bk4hJZt25crl3fwAD988TEduvYoXTfr2y8B+N8r75Kzby9P3nMr4557g7CwMPoMPI4TzzqXu668gGAT170vkWnN2Hz/DUS360zahWPY+kRl6cW+H78ib90qCA+n5fXjiOveh4OrrTSkEh1D8rDTyN/4R7ltijJ3suWRO4Iec9rxxxHfvi0zh5xJct/e9HzoXn4+q3KqToBmp55I8YHKzuWY5k1JHTKQg1vLsoJFJCXSc8I9/HrZv8jfvpOoJpUNYIeK086zE8+xXxpAl3gges0MVb1QVdNVtamqXlwht7jhEPGn2vRR3q9+U0SiReQHD0nH6yLSveaR1g/dWzZia9ZBtu/Lo9ilfL9qJ0O7pJUro0B8tPW7Mz4mgsz9BQC0T0tg4UZrMkPWwSL2FxTRrUUSBcUuFm2y8hoXu5S1O/aTnhh4Gs6q2LB2NektWpHevCURkZEcPewklvwyp1K5z99+jdPOu5TIqKjSdds3b6Rb3/4AJCWnEJeQwKY/LJFGx249SW6SGrQ4PYnv3Z/9v1qp7As2rSMsNt679nGdW/tYQsGWjeW0j03OvICsH75Ei4tqJcaKND3leLZ9amUt27dkOZFJiUSnVz4/4XGxtL/6MtY/+2ql17rfdwe/T3iqXMur5dmns/O7H8nfbqVFKNwTvAkxTjvPTjzH/hB1BbyEKv5MUHfYf58TkWcrLnUXYsOlFlSbfYFIVe2jqh+q6hhVXR3kY9QZ6YnR7MouKH2ekVNAWmL5wfKvzfqTU3s146ubhvDURX35v++sCm7drv0M7ZJOuAjNk2Po2jyJpknlK+aE6AiOOyK1tGIPBll7dpOS1rT0eUpqGlmZ5Uc+/7V+LXt3Z9Bn4OBy61t36MSSX+ZQUlLM7p3b2bRuLXt37wpabL6ISE6hOCuz9Hnxvj1EJPtu9YTFxhHfqx8H11qTFKJatSOicSoHVy6uVDaySTqt73yUljeOJ6Zj16DFHNMsnbztO0uf5+/YRUyzylKRI26/jg2vvUNJXnm/d/rJw8jfmcH+38u3VOPbtyWyURIDP3qD46Z+QMtzhgctZqedZyeeY7805MQpgPtmxWE6B6D2cacWtdOCjgcysQQei4BL7bSfpwJP269V/qSW7SsdeBdIs6Uc5wBvALep6m92StRngDOBPCyr1y4RGY51uyMKy8R2ib1+PJbQo4P992lVfdY+1uXAbVgN3OWqepmIpGFlUXNLQG5S1XnBOE+eVLwLdUrPZkxdtoPJ8/+iZ6tGjD+7Jxe99AtfLdlOu9R4Jl59DDuz81ixJZsSV9nW4SI8eE4vPvp1C9v3+XKcBCFAQKQsy6HL5eL9V55lzK33Vio35O9nsmPzX9x/3VU0SW9Kp+69CAsPD15sPvGShdHX/b6wMJqOupF9M7+leE8GiJB2zhXsmvRipaLFOVlsGvdvXAdyiW7dnmb/vJ3NE25F82t+vj3PaVnI5WNO6t6F+LZt+P3+x4ltVSb1CIuJodP1V/PrJZVy/yAR4TTq1Z0FF15NWEw0g6dMYt/i5RzY+FeNY3baeXbmOfZDSd30/tQm/uZhf2X/fdtXGUNQ6Qv0ALYD84DBIvIb8BpwArAe+NDXxqqaISJjsCroM6HSBy4emK+q94jIY8DVwP+AucBA+8fBGOAO4FZ7m67A8VjZ7daKyEtYc8nvAQaraqaIuJsIz2DlIp8rIm2AaUA3zwA89Zrtht9Iev8z/J6QjP0FNG1U1qJOT4ou7fJ2c1afltw42fods3JrNlERYSTHRZJ1sIinp5f9sn9t9AC27C27x3b3md3YsucgHyyouSTBk8apaeVaxXszd5frys7PO8i2TRt45I5rAcjeu5dn77uTG+5/lPZHdOOia24sLfu/m/5J05atgxqfm0ZD/07SsZ7ax1TAcrlEJDehODvL63bpF42laPdOsmd+A0BYdAxRzVvT8sb7AMso1XzsHex45TEKNm/AVWyZvAq2bKQ4cxdR6c0p2HxoY1bbXnEBUsKKSgAAIABJREFUrS86B4DsZauIbdEMd5QxzZtSsKt8T0ZyvyNp1Lsbx//8LRIRQXSTFAZ+9Aarxj1CXOuWDJn2cem2Q779kHnDLyZ/xy4K9+6jJC+Pkrw89i5YRGL3Iw65MnHaeXbiOQ4UDWELV6D4S036FX5MUap6Vq1EdPjyq6puBbBbyO2AXGCjqq6z17/LoQ/4KwTc978XUZYjvhXwoYg0x2plew5bnaqqBUCBiGQATbF+PHyiqplQmrYU4CSsfOPubZNEJNGdcMcu+yrwKsAxD3xf5ZDN37fl0DoljubJMezOKeDkHs347+flc8XszMlnQPsUpi7bQbvUeKIiwsk6WER0RBgikF/k4ugOKZS4tHSw2tjjO5IQE8GEr4J/t6B9l25kbNvK7p3badwkjV9n/sDYu8aXvh4Xn8BzH5fZaR+5/VouuPo62h/RjYL8fECJjoll1aJfCQ8PrzRYLVhkz55G9uxpVkw9+tJo6KnkLppHdDvf2seUMy8gLDaOjMkvl65z5eex8a4xpc9b3ngfmZ9PomDzBsISEi2HsyoRTdKJTGtOUeahd/H/9faH/PW29Zs1/YQhtB11EdunfEty394U799PQUZmufKbJ33E5kkfARDbqgUDJj7P/POtNPo/9B1WWu74n79l7hkXUZS1j13TZ9Djwf8g4eFIZCTJfXuz4fV3Dzlmp51nJ57jgHGVVF0mxPHXJe5OofkPoBlWdyvARcCmWozpcMWz6VhC2f8mWHMRijzm03vu/zngSVX90qNr3l9M4iOmMGCQH4d2tSlR5Ylv1/LsJUcRJsJXS7ezcfcB/jmsI79vz2HOH7t5dvof3D28Oxcd0xYFHpyyEoCU+CieueQoXKrs3l/A+C+s9emJ0Vw5pAMbd+fyzj8HAvDxwi18uWRbUGIOD4/gkmtv4f/+czMuVwlDTjmTlu068Pnbr9HuiK70HTTE57b792Xxf/fcjIjQuEkaV98xrvS1j15/gfkzplNYkM8tl4xg6KnDOfuyMT73VR0OrlpCXI+jaHvfs7iKCsl4t6zbtfVdj7HlkTsIT04h5dRzKNy5ldZ3PgpYruecX37yud/YTt1JOeN8KClB1UXGB6/hOhgc+2rGT3NIO2EIw+ZOpSQvn+W3/rf0teO++4i5p55/SPvNXb+R3TPnMWT6J6DK5vc/I3dtcKb9Oe08O/Ec+6UBVNiB6DVnq+rQqtYZqk+Fe9ieXdnPY40d+AD4AzheVf8UkfeBRHc5L/uruJ+ZeNzD9lBxngucqaqjRGQJMEZVF4nIW0B7VR3mRbG5Euv+dzyWQWyQqu4RkRRV3Ssik4Elqvq4Xb6PLSPxSiAt7FDiqcudl9jPialJ105ZU98hVIsuI4I3kK6ucNo5Bjhjy/Ia6y6Ll00P+Dsn4shTHKvXTBORDu4nItIeSPNT3hAkVDUfqwt8qojMxXJkB5vxwMciMgdrYFtVMa0CJgCzbI3nk/ZLN2CrOkVkNdb8fYPBYAgNXK7AlxAlkMQpNwMzRcQ9gqEdXtSKhurjbvFW1Gqq6nUej7/DGvwVyP4q7mdYxWPZjz8BPrEfTwGmeNnX+ArPe3o8fht4u8LrmUDws3oYDAZDENDiwqDty5698wwQDryuqo9UeD0aeAfohzX75gJV3VTT41ZZYavqdyLSmbJKY409EMlgMBgMBmcQpJaziIQDL2AN3N0KLBSRLyvkvLgKyFLVTiJyIfAoQWjQVFlhi0gccAvQVlWvFpHOItJFVX1m3DLULiIyGrixwup5qnptfcRzKIx77daqC4UQWa85836loXbJuO2lqguFGA+k+0znELL4nwAaGBq8QWdHA+tVdQOAiHwAjAA8K+wRlA3g/QR4XkTKibQOhUC6xN/CmgY0yH6+FfiYsilChjpGVd/C+r8Y6ghTWRsMDid4FXZLYIvH861ARQVyaRlVLRaRbKAJAYwT8kcgg846qupjQJF98Dy8puwxGAwGgyFEqcagMxH5p4j85rF45r/wVv9VbDkHUqbaBNLCLhSRWPfBRKQj5efnGgwGg8EQ0mhJ4C1szyRPXtgKeKYgbIWVodJbma0iEgE0AmosLQikwr4P+A5oLSLvAYOBUTU9sMEQKE508jrNe+y0eMF514XTPOlOdNH7JXijxBcCne0pztuAC4GLK5T5ErgC+AU4F/ippvevoYoKW6w8k2uwsp0NxGrm3+hOS2kw1DZOdPI6zXvstHjBedeF0zzpTnTRV0Wwconb96Svw/IlhANvquoqEXkA+E1Vv8QSL00SkfVYLesLg3Fsv/ew7V8EX6jqHlWdqqpfm8q6fgi2O9spONHJ6zTvsdPiBeddF07zpDvRRV8lrpLAlypQ1W9U9QhV7aiqE+x14+zKGlXNV9XzVLWTqh7tHlFeUwIZdDZfRAYE42CGQ6cW3NmOwIlOXqd5j50WLzjvunCaJ92JLvoqCWKFXV8Ecg/7eOAaEdkEHMCWP6hq79oMzFCeYLqz7f2Nx7fv+lKsVKNRwALg31h+7YGqeouI3Ih1a6SDPQjxbVU9TkQeAc4CioHpqnpbEN53pXWh7+R1lvfYefE68LpwpCe9PCHvoq8q/hBOORoogVTYp9V6FIbqUiN3tgfefNedsDLyDFbVIhF5EbgEmA7cbm83BNgjIi2B44A5thd7JNDV/vGQXOFY5XzY1yW35NQE7604Jzp5neY9dlq84Mzrwo1TPOlunOiir5IgpiatL/z5sGOwBA6dgBXAG6paXFeBGfwSLHe2N9/1iVj5bxfaLYBYIENVd4pIgogkYk1XmAwMxaq8PwNygHzgdRGZipfEOp5TJaa27u1zxKQTnbxO8x47LV5w5nXhximedDdOdNFXRXWmdYUq/lrYb2MlS5mD1cruTuV0mIb6IVjubF++67dV9W4v5X8BRmM1xeYAV2JlwLvVHjl5NFaFfyFwHVZrv0Y40cnrNO+x0+IF510XTvOkO9FFXyUNoEvcpw9bRFaoai/7cQRWq855UuAGQi24s8fj3Xcdh2XvGqyqGXZXd6Kq/iUio4AH7OUtYCWQp6pHiUgCEOexzXpV9TlyyV8LOxQxqUnrBqe5mhvPmVHfIVSbm99xXi7xBeNOrnF2zYMfPhzwd07cBXeHZDZPfy3s0vkZduupDsIxHCqqmm/fI54qIpnAXKxBadXdz2oRuReYLiJhWNfBtVgu7jlY3eGzVbVERLZgzdMH6z74FPtWimBpWQ0GgyEkCKL8o97wV2EfKSI59mMBYu3n7lHiSbUenaGUWnBnj6/w3NN3/SFeBq2p6p94DClW1VM8Hu/AstgYDAZDyNGgR4mrat3PGzAYDAaDoRZwFTp/zHQg07oMDqYhuLMNh4ZxNdc+7z3xr/oOodqMc9g4AQDGLa/xLhp0C9vQMDDubIPBYAAtMRW2wWAwGAwhj6mwDYY6wGkaRXCWrtJp2kdwrvrRSdcFOPOz5wtXA0+cYjDUO07TKIKzdJVO0z6Cc9WPTrouwJmfPX+4ipw/6CwQW5chBAm2btN+/ZCNYCKSe6jb+sNpGkVwlq7SadpHcK760UnXBTjzs+cPLXEFvIQqpsJ2KLWg2xwGhJzC02kaRXCWrtJp2kdwrvrRSdcFOPOz5w91uQJeQhXTJe5QgqnbFJF2WKKXEluteT1WBrOXsdSbADep6jw7BelzQH+shsz9qvqpvZ8JWOlN84ARqlrjb2/HaRStvVdeFaq6ygagfQSnqB8ddF3g1M+eb1wh3HIOFFNhNwxqpNtU1U0i8jLlc4tPBp5S1bki0gaYBnQD/gtke+SZb2zvJh6Yr6r3iMhjwNXA/zyPY/SalalPXSU4T/sIzlI/Ou26cOJnL1BCuas7UEyF3TAIlm7Tk5OA7h6/spNsteZJWDYuAFTV/XkupEypuQg4ueIOjV6zPPWtqwTnaR/BWepHp10XTvzsBYqpsA2hQrB0m56EAYNUtVy/mlg1uLf9FmlZf5lnDDXCaRpFcJau0mnaR3Cu+tFJ1wU487Pnj4YwStynXtMQ2tSCbvNWIElV77OfTwaWqOrj9vM+qrpURB4BYlT1Jnt9Y1XNcsdjrzsXOFNVR/mK3+g1ax8npiZ1mvrxvYzgzHeuS5ymMAU4Y8vyGusiN989OuDvnDYPvxWSekozSryBoqr5WF3gU0VkLpYe0x9fASNFZKmIDAFuAPqLyHIRWY01KA2s+9KNRWSliCwDjq+lt2AwGAxBw+VyBbyEKqZL3KHUgm7zD6B3hdWVsl6oai5wha947MefAJ8EclyDwWCoC8w9bIPBYDAYHICa1KQGp2F0mwaD4XAklBOiBIqpsA8zQkW32ea4VvUdQrV4bOh/6juEavPEjGfqO4Rq833tT+MOLq2dNxgxvZ/zYg4GrkLnjxI3FbbBYDAYGjwm05nBYDAYDA7AdIkbDHVA8yvGkthnAK7CAra+9CT5m/4s97pERdPmpruJSm8O6iJn0QJ2fTARgMjUdFqNvYnwpEaU5O5nywuPU7zXUiY2u/hKEvsOABFyVyxhx9uvBCXeHs0SuaBPS8JEmLtxD9+tySj3+vl9WtAlLRGAqAghMTqSm74oy9QVExHG/ad2Zem2bN63E3fcMKQDjWIjCRdYl3mAyYu3+kxDXV1Ulf/7YQnz/txBTGQ4951xNF2bVU4b++Ks5UxduYn9+UXMvvWc0vWFxSXc9/UC1uzMolFsFA+NOJYWyfEUlZTw0He/8fvOLMKAW086in5tK8sj6iPmxZszePLHJazPyGbCiEGc2NXqi9+RfYA7PptHiSrFLhcX9OvMOX07hUTMdX2enXiO/b6fBtDCNvOwDSFNYp/+RDdryR83j2Hba8/S8qrrvJbL/Poz1t02lvV3XU98l+4kHGlpH5tfchVZc35k/Z3XkvHZ+zS7cDQAcZ27EXdEd9bdcS3rbv83cR2OIL5brxrHKwIXH9WKZ+ds4L5paxjQpjHNk8qbpD5aup0Hv1/Lg9+v5ad1mSzZVj495Yiezfljd/lMVa/+sokHp69l/LS1JEZH0L9VeS1jTfh5ww42Z+3ns7Gn859T+/PItEVeyw3p1IK3r6iUcZYpyzeQFBPF59ecwcUDuvDczGUAfL7UymX9wVWn8vyFw3j6p6W4gvQro6YxN0uK574zjuHv3duUW5+aEMMbl53I5Cv/zsTLT+LtX35n9/7gSECcdp6deI79oSUa8BKq1FqFXQu+5mgR+cFO7HGBiLwuIt1rHmn9ISLXiMjlVZSZKSL96yqmmh5XRG4SkbhgxZHYbyBZc34EIG/9WsLj4olIblyujBYWcGD1cutxSTF5G/8ksonlEI5u1YbclUsBOLBqGUn9Brq3IiwyEomIQCIjISKC4uzKeZ2rS/uUODJyC8g8UEiJS1m4OYsjWzTyWf7oNo35dXOZAKJN41iSYiJYvWt/uXL5xVbrIFwgPExqlHO2IrPWbeOMnu0QEXq1TGV/QRGZuZW/QHu1TCU1IbbS+tnrtnNGr3YAnNC1FQv/2oWqsnFPDgPaWurOlPgYEmIi+X1HcJSVNY25RXI8ndOTKxmpIsPDiYqwjGOFJS6C2SZz2nl24jn2R0lRScBLqFJrFXYt+Jr7ApGq2kdVP1TVMaoavCz99YCqvqyq79R3HEHmJiBoFXZkSipFe8oMQUV7M4lMSfVZPiwunqSjjiZ3pdX6yP9rI42OPg6ApAHHEh4XR3hCIgfXrSF39XK6vfQu3V56l9xliyjYvqXG8SbHRrL3YFHp8315RTSOjfRaNiUuktT4KNZkWKYlAc47siWfLNvutfyNQzvwxIie5Be5WLS15j8u3Ozen0fTxLJ/WXpiLBnVaPFk7D9Yun1EWBgJ0ZFk5xXSOT2Z2eu2UexysW1fLmt2ZrEr52AVe6ubmP2xM+cgF73xHWe+8BWXH9OVtMTKldGh4LTz7MRz7A9XiQa8hCq12cLOtf8Os1trn4jIGhF5zxZIICKn2uvmAv/ws6904F2gj93C7ujZAhSRXBGZICLLRGS+iDS11w8XkQUissRunbvXjxeRN+19bBCRGzyOdbmdjnOZiEyy16WJyKcistBeBvuJ9VkRGWc//ruIzBYRr+fZjuM2+3EfO/blIvK5h7YS4FIR+dlOB3q0n2MfbZdbYv/tYq8fJSKfich3IrLO1l+6t3lJRH4TkVUicr+XfV4lIk95PL9aRJ4UkXgRmWqfp5V2r8cNQAtghojM8BVntfCqEPbtEG5z/Z1kTvuSooydAOx473Xiu/Wk08PPEd+tF0V7MtGSEqKaNie6ZWvWXHs5a/59GQk9jiSua8/aCNdna3hAm8Ys2rqv9F70sE6prNyRQ1Zekdfyz8zewO1friIyXOianuC1zKHgLT4vKuRqbY/AWb3bk54Yx+UTv+fJH5bQu2Uq4WHB+cqpacz+aJYUx/tXncrnY89g6spN7DmQH5T9Ou08O/Ec+0NLXAEvoUpdDTqrqa85Q0TGUF5y4VnEl4t5LjBQVdXe/g7gVnubrlh5sBOBtSLyEnAEcA8wWFUzRcQ9wuIZvLuhvXEXsFBE5gDPAqeraiBXwDvA9ao6S0QeAO7Daq0CxKvqsSIyFHgT8FWzrAGGqmqxiJwEPAS4R4H0wfo/FNjv9zlV3QLco6p7RSQc+FFEeqvqco99fgAsF5E7VLUIGA2MBU4FtqvqGQAi0khVs0XkFizhSHkPH+V92OP69+DcTm0qFgEg5eQzSTnh7wDkbVhHZJO00tciU1IpztrjdbuWV99Awc5t7Pl2Sum64qy9bH5qAmA5hRsdPRhX3kGSTjyNvHVrcRVYXxT7l/1GXOeuHFyz0vuZDZCsvCJS4spa1MmxkezzUQEPaJ3M5MVbS593aBJH59QE/tYplZiIMMLDhPxiF5+v2FH2flzKsm3Z9GnRiN935R5ynB8tWscXy6x7n92bp7Brf1mLLGN/Hmleujh90TQxjl37D9I0KY5il4vcgiIaxUQhItxyUt/ScldO+oHWKYf+QyOYMQdCWmIsHVKTWLpld+mAqeritPPsxHMcKKF8bzpQ6qrCrg1fsye+XMytgA9FpDkQBWz02GaqqhYABSKSATTF+vHwibuyUVX3jSCvbmhVLX+j0drmoIhcDcwGblbVPyuWqYiINAKSVXWWvept4GOPIu/b+54tIkkikqyq3vpEGwFvi0hnrB/Inn2xP6pqtn281UBbYAtwvl2RRgDNge5AaYWtqgdE5CfgTBH5Heu2xAoRKQCeEJFHga9VdU5V79PTh73iotN9fnr2fv81e7+3/p2JfQfQ5JThZP88i9hOXSg5eIDifVmVtml6/uWEx8az7dXyyULCE5Moyd0PqqSNOJ+9M6cDUJS52/pRMCUMRIjv1ovMb7+o6i1Uyaa9B0lPiKZJfBT78ooY0KYxr8+v7F1pmhhNXFQEG/aUfSG+sWBz6eNB7VJo1ziWz1fsIDoijJiIMLLziwkT6Nk8ifWZNdMnnt+vM+f36wzA3PXb+WjxOk7p1oaV2/eQEB3p9Z6kL4Z0asHUFZvo3TKVn9ZsZUDbpogI+UXFqEJsVAQLNu4kQsLokOr7fn5dxuyLXTkHaRQbRUxkBDn5hSzfmsklA7qERMx1cZ6deI4DJZS7ugOlrirs2vA1e+LLxfwc8KSqfimWhnJ8FTH5cj17dUP7oRewB6t7OBhUjMnXeXsQmKGqI0WkHR5SELy8XxFpD9wGDLAVmROBGC/7fR34D1YL/i2wZCEi0g84HXhYRKar6gPVeVOBsH/JQhL7DOCIp99ACwrY+kpp7zydHn6O9XdfT0RKE9JHXkj+ts10euhZAPZM/5qsGdOI79aLZheOAuDA7yvZ/tYLAGQvmEt8j950fuxFUMhdtoj9i3+tcbwuhfcXb+WmoR0IE2Hexr3syMnnrB7N+CvrIMu25wDWYLOFmyv/8PBGVHgY1x7XgYgwIUxgTUYus/6s1IFxyAzu2Jx5G3Yw8pWpxERGMO70srsuF785jclXWr0dz85YxrTVf5FfVMwZL3zJiN4d+OeQnow4sgP3fTWfkS9PJSk2igkjBgGw90AB1380izCBtIQ47h9+TMjEvGrHHu74bB45+YXMXb+dV+au5KMxp7FpTw5P/7S09IvgkmO60ik9OCPynXaenXiO/RHKXd2BUms+bAm+r7nifmbaz38THy5mEVkCjFHVRSLyFtBeVYeJyHggV1WfsLdZCZyJ1bX+OVblvEdEUuzuYq9uaB9xtgW+B4YB3wBjVXWBj7KlcYilqrxOVefY6xup6s32+1yjqteIyHHAS6rqdf6RiHwOvKuqn9r7GKWq7URkFNDfbfISazT+E0AWVld8XyANq2V9p6pO9Dy/9jaL7TK97cq9BbBXVfNF5Gz7WGeLyArgLFX17M2ohL8Wdijy3MiH6juEavPEgc/qOwSDISgkjX6gxnfPF5x8fMDfOcd8PyMkfdj1ljjF/qJ3+5ozse4313zUT3nGAx+LyDZgPtC+iphWicgEYJaIlABLgFFYbugXRGQ51jmbTZkfuhSx+szfwKrotovIVcBEERlg+6m9Htb+ewXwslhTojZg3St2kyXWNLkk4Eo/b+ExrC7xW4Cf/L1X+/0us3/UrLKPOc9P8Y+APqrqbhb2Ah4XERdQBPzLXv8q8K2I7FBV48o2GAwhQUNITVprLWxD1YjIc8BiW8gR0tit8qdU9cdg7M+0sGsf08I2NBSC0cL++W9DA/7OOXbW7JBsYZtMZ/WEiDwIHAN8Wd+x+ENEkkXkDyAvWJW1wWAw1DV1lelMRFJE5Ht7Cu33Un6KbsWySSKyzb5VXCUhl0tcHOJrrk6cInIPcF6F1R+rqs851cE6dk2xR6MfEez9tj2lzpO31QjTWjV447Z4n+kjQpbrP3eeKrbX6KrLVEUddonfhTUr5xERuct+fqePsg8Cs3y8VomQq7BDxddcFdWJU1UnABPq49gGg8FgAHXV2V24EViDjsGaojsTLxW2PcumKfAdEFALJuQqbIPBYDAYgk1JYeA5wj2TPNm8aueRCISmqroDQFV3iJWps+L+w4D/Ay4DTgw0LlNhG0IaJyr+DjeNotFrBobTtKvgPLWtP6pzb9ozyZM3ROQHoJmXl+4J8BD/Br5R1S1SjXyvZtCZIaRxouLvcNMoGr1m1ThRu+o0tW1VBFP+oaonqWpPL8sUYJedXRP7b4aXXQwCrhORTVg5MS4XkUeqOm5IVNgSZBWnUxCRB+yc3/7KbBIR33qqWuJQjysiQR3R4kTF3+GmUTR6zapxonbVaWrbqlCXK+ClhnyJlVcD+++UigVU9RJVbaOq7bCyTb6jqndVteOQqLBrQcXpCFR1nKr+UN9xBJmgVthOVPwdbhpFo9esGidqV52mtq2KOtRrPgKcLCLrsLwWjwCISH8Reb0mOw6JCluCqOK0y/rTZ14qIr+Kpel8RUTCReR8EXnSfv1GEdlgP+5oHw8ReUREVoulv3zCz7GniMjl9uOxIvKen7IT7VSqiMiJYmkxV9ixe/aX3W7H/KuI+LyhJoemE/1CRBaJpdesJF8RkQdF5EaP5xNE5AYRaS6WOnSpWHrNIXaXTqy9zuf7rg5OVPwdbhpFo9esGidqV52mtq0KV2FJwEtNUNU9qnqiqna2/+611/+mqmO8lJ/oThtdFaE46KxGKk4PvOkzOwEXYOkzi0TkReASYDpwu73dEGCPiLQEjgPmiKXZHAl0VVUVEX83iv4JzBORjVgqz4F+ygIgIjHAROBEW6rxDlaqz6ftIjmqerT9Q+BprLzn3qiWTtTWZV5p50uPxdKCfqqqnv7KN4DPgGfskY0XAkdjpWydpqoTxFJzxtl50K9T1T4+3mfpyMunLz+D0cOO8vomnKj4O5w1ikavWTVO0a46WW1bFcbWVTsES8XpTZ95ItAPq2ICiAUyVHWniCSISCLQGpgMDMWqvD8DcoB84HURmUqZyrMSqrpLRMYBM4CRHopOf3Sx398f9vO3gWspq7Df9/j7FL6prk50K3CDiIy0y7QGOmOZxtzvZ5OI7BGRvvY2S2wxykLgTRGJBL7wJUPxxHPkZc5b43x+epyo+DucNYpGr1k1TtGuOlltWxXGh107BEvF6Uuf+baq3u2l/C9Ywo21wBwsycYg4FZVLRaRo7Eq/AuB67Ba+76orl6zqs489fG4ItXSidplTsKykx0Uy9DlS685Cmsaw5tQ6uYeCpwBTBKRx1X1nSreR7VxouLvcNMoGr1m1ThRu+o0tW1VBGuGQn0SEvIPCb6Kczze9ZlxWCP2Bqtqht3Vnaiqf4mloHzAXt4CVmLlzz5KRBKwunzd26xX1cqTPq1jHY3VijwdK+XcKb5Uk2L5p7+2lz+AE1R1vb1+iao+I9aw/5ftNHeXAheo6nAf+6uuTvRIu/xwEekKLAVOVdWZ9nH7q2qmiEQBK4BIoLOqloilEd1m/5i5CWinqjeJSBaQbne3+8RfC9tgcAomNWnd0Ov9b2o8QuHzZj0C/s4ZuXNVSMo/QrGFXYlgqThVdbWI3AtMt+/HFmF1Pf+F1apuDcy2K6QtwBp700Rgin2vWYCbve3fHij2GjDa1mveitVtfIL6/mWk9vsbjaUCjQAWAi97lIkWkQVYgwQv8vMWx1MNnShWSrxrxNKGrrW38RZgoYjMAPapqntExjCswXBFWLcsLrfXvwosF5HFqnpJFcc3GAyGOqEB9IiHRgv7cEVEvsLqwp5R37H4w/5xsxg4zz2OoKaYFrahIWBa2HVDMFrYk9O6B/ydc/Hu1SHZwg6JaV2HIyLyJlYX/dz6jsUfItIda2T+j8GqrA0Gg6GuKVENeAlVHNEl7gupRxWn+FZmVrJyicgLwOAKq59R1Str+9g1RVVXAx2CvV+DwWCoS0yXuMFwiJgu8drHdNfWPk7zujuVpNEP1LiL+o2UrgF/51y1d01Idok7uoVtMBgMBkMgNIQWtqmwDQaDwdBYxrQ4AAAdIElEQVTgMRW2wVDLONF77DS/tBM9zeA8V7PTrgsnfvb8Uehyfo1tRokbQhqneY+DEXNd+qWd6GkGZ7qanXRdBCPe+vjs+aMhjBIPyQpbguzHtl8/ZIWn2DaxYCPGh10lTvMeByPmuvRLO9HTDM50NTvpughGvPXx2fNHiQa+hCohWWHXgh97GBByzm3jw64ap3mPwVl+aSd6msGZrmYnXRfBiNcftfXZ84dpYdcSEkQ/toi0A64BbrY9zUNEJE1EPhWRhfYy2C6bICJvieWkXi4i53jsZ4KILBOR+WJ7pn0cz/iwjQ+7EqHql3akpxkc6Wp20nXh63ih/tnzR0NoYTth0FmN/Ni2HvJlyssvJgNPqepcEWkDTAO6Af8FslW1l13O3ccWD8xX1XtE5DHgauB/Pg5pfNjGh+0IvzQ4x9MMznQ1O+26cOJnL1CKQrjlHChOqLCD5cf25CSgu8e9lSSxXNgnYVVIAKiq+2ZdIWUO7EVA5REWZdsYH7YPjA87MOrKLw3O8TSDM13NTrsunPjZC5RQ7uoOFCdU2MHyY3sShuWALndDxu5u97bfIg/blmcMvjA+7CDhNO9xMGKuS7+0Ez3N4ExXs5Oui2DEWx+fPX+Ecld3oIRkalIJvh/7ViBJVe+zn0/Gaik+bj/vo6pL7XuwMap6k72+sapmueOx150LnKmqo3wcy/iwjQ87JDCpSWsfk5q0bghGatI7ozoE/J3zaOGGkExNGpKDzqpCVfOxusCn2oPOKvfhlecrYKR70BlwA9DfHli2GmtQGlj3pRvbg6iWYd3zDRgp82Ffqarbse4fvynid6iG2u/H7cNeAbjw7sO+ER8ubpvx9j7mAIE0ab7DamkvBx7Ejw8bq4v/owo+7KX2j4RzAHefo9uHHZRBZwaDwRAMGsKgs5BsYR8uiPFhG2oR08KufUwLu24IRgv75oj2AX/nPFW8MSRb2E64h90gEWf5sL8GPjc+bIPB4FQaQmrSBlVhSx36scX4sA0Gg8ExhHJXd6CYLnFDg0NE/mlPIXMETosXTMx1gdPiBWfG7CQcOejMYKiC6szJDwWcFi+YmOsCp8ULzozZMZgK22AwGAwGB2AqbIPBYDAYHICpsA0NEafdQ3NavGBirgucFi84M2bHYAadGQwGg8HgAEwL22AwGAwGB2AqbEOtIiIjRUTtXOX1GccwETm2Btsni8i/RaTETnHrXtqJSH8ReTaY8VYRy88BlMkVkWYi8oGI/Ckiq0XkGxE5oi5irBBLtdOXicg9Yjnal9vn2afRQiwX+yq7XDcRubhmEQcco/taWCYii2tyfVVxnFG2RyEY+yoRkc0iskZEvhKRWrdu2PEHKkIy+MFU2Iba5iKsbG4XVlWwlhkG1OQLNRn4N5Cnqn08lk2q+puq3hCUKANAVQN9H58DM1W1o6p2B/6DpUita6pVYYvIICw5zVGq2hvLKLfFzyaXAE/YHvamQFArbLHw9l3pvhaOBO4GHvaybXgwYwkCecAG4FJgL5bGt9aw3/8oAjcXurdrUEm9goaqmsUstbIACcA24Ahgjcf6O7DsX8uAR+x1nYAf7HWLgY5Y2tHHgZV2+QvsssOArz329zwwyn68Cbjf3scKoCuWQ32nHctSYAgwHFgALLGP29TefjyWQnQm1hfbDfb6D7C+7ErsmJoDs+39bQR+9re9/doXWD71VcA/PdbnAhPs9z7fI5amWJXuMns51l3e4/z+6PFeR3jsMw+Y7eV/4u+czgI+wjLHPYJVEf5ql+tol5uIJaaZY5c7014/Cnje4zhf2/t8xD5nS4H37Ncutfe7FHgFCPc4Dw/Y+53nJfYT7f/XCvscRwNjsCqejcB79vnLtvd9M/AN0Nvefgkwzn78oL2t13OIdc38Drxob9cWOAX4xS77sfv/YJc/D8sL7z6XM4DJwGp73S32OV8J3BTANTHaPg+zsIRCz1dxTVTaDxBu/7/c/+ubsa6LXGAt1o+gVzyOeTuwEFgO3O9xHtYAb9vrPwHifP0/PD6D47B+qF/qcbylQCzQz35fi4BpQHN7u5nAQ/Zrt9rndKX9Pitdy4fjUu8BmKXhLvaH9Q378c/AUcBp9mP3hz7F/rsAGGk/jsHKs34O8L39xdMU2IxVUQ7Df4V9vf3438Dr9uPxWKpW9zaNKRt0OQb4P49yP2NVBqlYXvNI+4trJWWVzzZglb3NCcC3/rav8F5j7X01sZ8rMNx+/Bhwr/34Q+wvd/scNLIfuyvsCCxtLPax1nu8pwLgKS//E3/ndJ/9ONp+f+4v7RuBp+3HE7Esb2FAZ2Cr/f8ahZcK2zNe+3E3LHue+5y8CFzucR7Ox6pEl2JVWC8Cf7OPsQU4wi77jse5mQicaz8eRvlr4y6sVmQSVmU0zV4/A+ji6xza/28XMNDjtdlAvP38Tvv1pVgVWjbQzyOGA1h6W7AqqBVAvP3eVgF9fV0T9v9gM5AGRAHzKKuwfV0T3vbTD/je41wkY1WeM4GjsX50nGq/dgrWCG+x/7dfA0Pt86DAYLvcm8BtVfw/NgF3eBx3JpaqF6zP0s9Amv38AuBNj3Ivemy3Amjpjr2+v89CYTFd4oba5CKslin234v+v70zD7OrKPPw+yMICQTjIIh5MJAQEWTRqEFBorKLywBqGBfQBNmCgyIKgwuDDDIiExwYiMqmhH1YAhhAIYiGYAghIWQBJDCETQQJYZGEBCF888f3nfTpy7m3u9PddN/09z7Pfe65depU1amqc+vU9vvw4c0LzOxlADN7TtIG+IN5bbitiPOjgMvNbKWZ/Q1/896hHfFeE9934384VbwLuDnMmR4LbFs6d6OZvWJmzwLP0HoYebn50OuXcbOnJ+Ja6yvbcf23wmzrncAQvMED+Af+B1mb5t2AXwJEHrxYcw8CfhLmUX8PbErbQ96N8nSWmT1lZq8ADwNTwn0BrfPxSjN73dwYzCJ8FKO97I43JLMkzY3fhVb9SmCSmS0NP4cBi/FG6nDgETN7MPxeiDcobXF7+BsF3AgMlLQebr99IY3z8DEzK0zO7ghsA0yPdI8BVpoPiW8N7A1cVDKle5eZPRLHo3DjOcvi3q7BR3mguk58BJ/KWGxu2vaK0v3UqxNV4SwCtpB0lqS9gb/jDfpIfFRhQ/zlDbzB3gvvMc/By7Son0+Y2fQ4viTuZysal0c5zWW2ArYDbol8PB5/Fquumw5MlHQo/nLS58l5gqRbkPR2/M9lO0mGP3AGTIrvVt7rBVPH/TVar7/oX3P+lfheSf06fhZu2nSypF3wnnHt9XXDMLNpkj4OfAafoy03pm+4PuLYA9jJzF6WNLWU7lctuhFtpLmWA/Be2IfM7FVJj5bCfB1v9GppZDawnO7XS79fr0lTbfkZbZdJOf4Lzez7FedWWNhbj++pwNR4qRrTIN2NmIU3UIvwxmkj4FD8xQga5+GymnTfYmZfXuUgLS2OzWyGpI0irKpr30AbdaLd+23rhWNmz0t6P/BJfJThX/Ah8dnAj4CTw/3MSOMpZnZOTdhDK9Ji9e6pxLI67sJHpnZq6zozGxeLDT8DzJU0wsyWtBHvGk32sJPuYjRwkZltbmZDzWwIPs/4HPD16OUgaUMz+zvwF0n7hdu6cX4a8EVJ/SRtjL/B3wU8BmwT/gbhvbS2eAnYoPR7ED7sC+1rDFpdL2lz4BkzOw+fJx3UxvWDgOfjD3VrvMfWFrcCR0R8/SS9tSLMZ6Kh2RWfZy1YiY8AHFpK8w7A81TnaUfYX9JakobjveOF+DDoiHAfgg+5Frwq6S2lexot6R2Rpg0jL1chaStJW5acRgB/A4ZKene4fRUfHailVTlFD/UJvLG6E+9xHxPf0DgPy9wJ7FzEH/VzVaMVZdoPnwKpZRqwn6T1JK0PfC7ir1cnZgK7SHp75FvZMl9VnagMJ14g1jKzScC/41NSRR6tBXwLOCbiuBl/LgfGtZsWZQRsFgsBoWUR6QO0rzyK+IoyWQhsXIQn6S2Stq26SNJwM5tpZicAz+IjB32a7GEn3cWX8QVHZSbhc5iTgdmS/oE3dj/AH/hzJJ0EvIr/SV0L7IQvOjF8XuxpAElX4otgHsKH8drieuBqSfsC38R71FdJehL/Mx7W6GIzWyJpOt7YFYu2jpX0Kv5H/Uij6/F533Ex9Low4myLo4BzJR2MN8BH4IueCi4Frpc0m5a51DKfA86Q9D1gBd6ofhufR22Vp+rYtruF+J/zJsA4M1sRefMIPnx+Lz6sWnAuMF/SHDM7QNLxwJRYef0q3st7rOR/IHBWbDl6DZ9XPgy4HC+ztfGe89kVaZsPvBbDwxPN7HS8cdw9GrTb8SHYosFuKw8BMLPFksYCl0taN5z7x7AueOM9xsxWtoyKr7p2jqSJtLwYnW9m90i6n4o6YWZPxVTLDOCpyMtiSLiqTtSrW5sCF5RWuH8fn7eeGHm3HC+vL5nZxZLeC8yI9C/F16CsxBffjZF0Dv68/TLK/CDaLg+K+CQtx5/n0cCZ8bK9NnAGPq9fy/h4cRP+ojKvTvh9hlQ6S5Kk3UTDc4OZXd3TaUm6nxgSv8HMtuvhpCTkkHiSJEmSNAXZw06SJEmSJiB72Emforyyt5PhdFpyVdJESaNX47oOSz2qA1KfXRVnG+EVsp73yaU9v6NqNbHOxDFU0r1dGWZniPrys9LvY2Kuuk/QVc9eXyYb7CRZPXpScnUsHZB6VMelPjsdZzsoZD23BfYEPo1vNWoq1DHp0VeAz8fq7STpMNlgJ02HpOsk3R29s8PCbamk/4ze2p2SNgn3YZJmSJol6cddFP9AYGfgYKLBlhsXmSrparlhhUsVy20lnRDx3yvp3MK9FN7ukq4t/d5T0jWxbWdiXLdA0tHRIx8JXBo91AGSfio37jFf0mkVSR4MPBuCKISgy3s7GeeHJN0W5XCzpMERzlRJp0uaJunPknaIcB+SdHL4WR9fYT0vesC74qvAj5TTT9L4yLP5kg6P666Q9OlSmidK+kI9/zV53F/SBXFP98i3cBUjB7+RdJOkhZJ+VLrmQEl3xT2fUzTOUddOkjQT2Kkd+V/wGr5i/uiK9G0saVLcwyxJO4f7ArnhGUlaIulr4X6xpD0kbVtK43y13g7X5aiHn70+T09LreUnPx39UC3DWE/eczIt0pf/SkkmsxPxV0mu7oKLp7wLfxGeAYwqpzeOLy6lcyK+xUX4dqJCrvEyXOv8DdKS8T2VFqnHDfGtPCr7qUlvldRnZ+JsS17y1Dg+CvgrLXKnf4my+gIuFlPEUchrPo9vFTusVH7r4kIfw/BtaheG+zr4KMGABv6HAveG+3dxhT1wFa/HaZFUfSrSVdSnkbRDQrW9+V+6z6W4ROqj+N7pY4ATS/lf1JfNgD/H8dm4cMh2+Nap88L9oSjXs4ADSnkyYE1+9vr6J3vYSTNSJcNYT95zZ3z/Lnhj2RVUSa6Cy1H+xcwKjekiDbtKmilX7NqN1jKomP+jXQwcKN97vBPwO6qlJWv5O77H+nxJnwdervVg1VKfYzoRZ1vykpPjewGualXInS7Cy2sB0E/SqZI+Zi3ymsXIw17A1yLsmXijsGWkbzf5PuhP4QYhljfwX2ZU3C9m9gC+77swNXqLmS2JsK4Jv21KqMZxm/lfxlwk6CJctKTMHsCEiGsy8Fa5ZG8hrfpxXJJ0e0mbAs9Fuc4AfiDpOGDzuIfupKefvT5NCqckTYXqyzk2kvfssq0Qqi+5+luqJUn7472zkWb2hHyRUZVs5wV4j24FcJWZvQZUSUt+vXyRmb0m6cN4g/Il4MhIHzX+qqQ+D1+dOGlbXrIsaVord7q2mT0o6WW84T5F0hRco3olrr0u3IDLzbUBR3l/Eu/VF41BpX/5HuJymutRT3qzPRKq7cr/Gs7AxVAuKLmthdfpVg2upGl4OWwG/BAfZRhNCL+Y2WUxNP8ZXBv/EDP7QxvxrxY9/ewlOYedNB8dlficTsvCsAO6IP56kquj6vgvGudn5XPflavCzeyv+PDx8fhQeVvSkhuEn4H4kPJvcRWzEbVhq1rq87HVjZMOyEtWoVhtbmaXAKfhZXg2bpHKcJnMIxRyppLeI5/3Bh/ROAg3nlE00I38F0wjyl/Se/AGcGGc21MukToA2A+vM21KqIZ7Zf7LdxG8wT523PdzuBnTg0vOU/DGvgh3RPh9Atc/39LMFuELHVdJq0raAlhkZmfiPfP3VcXZRfT0s9fnyR520mx0VOLzKOAySUfRMozZGepJrh6BW7hqhZm9IOk8vDf5KD4PWY9L8Xnh++N3lbQktJZ6/BTwm+jJi4oFTdSX+lzdODsiL1nF9sD6ERb4XPS5wH/H7/PxYdU5koQP4+8X56bgQ8qTzXXC2/Jf8ItI/4LIg7Fm9op750/4kO27gcvMbDaA2pZQBX+Jqcr/4VRPJxT8jFIDjQ+R/zzq9dr4C8a4ODeTFmnS24FTIs3gIw0HyiVyn8btiXcXPf3s9XlSOCVJegmSJgD3mNmv1uQ4exNyffCRZnZkW347GO4lwNFmtrgrw036NtnDTpJegKS7cdOC312T4+wrmNmBPZ2GZM0je9hJkiRJ0gTkorMkSZIkaQKywU56BVpNbW65wtgNcbyP3PYzkvaTtE13pLVZWd08rgmjS/TPuzotks5fnfKWNEKt1dNW1aHVTNObkseS3inpfyU9LFdZ+62k49SFeu9vNlFHJqzmtW+T9I2uTlNvIxvspLfQaW1uM5tsZsUK7v2ADv2BS1rT13T0Jv3zLk2LmR1SWuneEUbgOuZFOOU6tDp0ex7HSvhrgalmNtzMtgF+EHE2bLBVo30uZ01oB94GZIOdJN2N6mtz31DyMyFW9CJpb7le95+Az5f8jA1/HwX2AcbLNZaHR0/qTrne8rWS/imumSrpJ5JuA46StL9cR3ueXLRijaBBHveE/vlG+Ha0wcD3JJ0WaZkWZXO/pLOLhkSuVf0zSXMk3Spp44r7myppZBzvHX7nSbo13D4s6Q65jvgd8r3p6+DboL4Y6fpiuZcnafOIb358bxbuEyWdGeEsKvXyhwP74vumvy/pY92Rx7hIyzBce32BpKPxLWlb4tvQlkeY10VePizpcUmXAQvkVsz+LOkXuIDLEEl7yXW/50i6KuoLqtBJj/s/W9Ltkh6U9Nlwb6TXfo1cr/0hSf9VuqeDIozbon4W7vW01U+U9OvI00WSCsW4nwLDoxzHSxoc9Wlu5MXHautMU9LT2qj5yQ/1tblvKPmZgPfS+uP7drfE971eWfiL8xPieCIwunT9fOATcXwScEYcTwV+UfK3ANg0juvqQjfbp0Ee94T++eHACxHGHbjs5i644toW+J7jW4ryw9WyCr3sE6rKuAgf2Djqx7DyfeAa3mvH8R7ApNo6U1GHrgfGxPHXgetK8V4VebYN8H/hfikwu5THo7opj8fjwje1eXwvLXu2t8L3jg+ONCwr5clQXHVux/i9Eb7ve/34fVzkc6VOeqTvprifLXGN+P401mtfhAuv9Mf3sg+JtD0eZbYOLrQyoXSvVdrqJ0berhvpXoJr2w8ldOPD33eBH8ZxP2CDnn4Gu+KTPeykN1BPm7uKrYFHzOwh86fxkrYCl4t7vM3MbgunC/FGouCK0vF0YKKkQ2kRq1gT6E365/vif7TnA/fH7yIti8xlPy+nRT3udVrK6BLqq8qBq29NM7NHIp3Phfsg4Cq5dbDTa++nDjvhDQdxr+V4rzOz182H4TcJty2AwXL52dtr7qsr8/hZYFBFHg8Cbo78exhXp7sIV4WbVeRJ8JiZFcInO+IvHtPlWuZjgM1prJN+Zdz/Q3iZb01jvfZbzexFM1uBl/nmwEfwYf3F5iI45eewnrY6wI1m9oq51blnSvlfZhZwUJTF9mb2UoWfpmNNn7NLejmqr809mdZTNmX97a7ei7hsVcBm4yR9BNdmnitphJkt6eL43lQa5PGbrn8eadkV/6P9LK7C1g83HlGl6V1Fo/JXnfM/Bv5oZp+Ta4xPbRBGPcrhlvNNcV8j8BeRI/Ee/Ut0Qx5Lmo33pqfSOo9XDamHv9/hqnajcGMtZZaVjoWPirzhRVn1ddLr6a/X4w15UCecgnra6o3CakmM2TRJH8ef44sljTezixqkrynIHnbS09TT5gbYRtK60UPePdweAIbFfCHU742v0r42twb1fGke66vAbVUXSRpuZjPN7AS8JzOkMzfXS+hN+uej8V7rDma2CZ6//SItH5bbUF4Ll9ws5DfXKqXhKyX3KmYAn5A0LNKzYbgPAp6M47El/2WN9FruoLUWdqN4R+OWvoaa2UbAvwHL6YY8BubF9RsSeSxpB7wR/pR8HcFQWix8TcBfIOpxJ7CzpHcDSFpPrsfeSKd+f0lrxXO4BT503kivvYqZwC6S3i7Xgd+/dK5SW70BrcpRrvv+jJmdB/yKlnrY1GQPO+lp6mlzfwWfn56P2/69B8DMVkg6DLhR0rP4n2ht7wF82Pe8WJQyGh/mO1vSevgQ3kF10jNebihDuAGIeZ24t95Cr9E/j/OHAjeoRX/7Cry8Z0Q6t8f//IsFV8uAbeXKbC/ijXklZrY46sc1kYZngD1xO80XSvoOULZm9Ud84dtcXKO7zLeAX0s6Ftcnr1dnwPN4Nj4q8ypu+/r8uK/uyON1457WwzXET8RtY58UcT+K90Sn4KMYdVfQR56NBS6Xmy4Ff0F4ifo69Qvxl95NgHHxXDbSa6+K96kYWZiB2ySfQ8s0VCNt9aqwlkiaHlMev8NHII4tlcXX6l3bTKTSWZIkq426SItcbrrxGDP7bMW5pWY2sDPhNzNdlcddhaSJ+ELPq3s6LX2N7GEnSbJaKLXIu53M46RM9rCTJEmSpAnIRWdJkiRJ0gRkg50kSZIkTUA22EmSJEnSBGSDnSRJK2Jf7Nz4PC3pydLvdbopzg+Gale98ztK+pOkhXJN7nMlDZB0iKQzuiNNSdLbyFXiSZK0IpTdRoAbWwCWmtlp7b1eUr+Qx+wIH8T3099UEd5gfK/2/mZ2V+yv3h/fX5wkfYbsYSdJ0m4kXS/pbkn3STok3NaW9IKkkyXdhSuW7RO94dtD8/q68DtQbu3pLrlFp3+WNAA3NnFA9OJrVb++iRsuuQsgNKyvMLPFNWnbV67NfY+kKZLeEe67yS13zZVbo1pf0qbRYy+sOX20m7MuSTpN9rCTJOkIY8zsuVCMmy1pEq6INQiYY2bHx7kHcXOJj+OKdQUnADeZ2Vi5idOZwPtwha7tzOzbFXFuB5zTjrRNAyabmUkah+9dPg44FjjMzGaG3OYK3HrZ9WZ2qtxG9ICOZkSSvNlkg50kSUc4WtI+cfwuYDhugeoftEiJbgMsNLPHACRdTos05F643vX34nd/XHO6K9gMuFLSO3HpzgfDfTpwhtwe9CQzWyppFnBOyG5eZ2ZrggRtsoaTQ+JJkrQLSXvgBiV2NLP34zrvhRGL5daiwtTIapOA/cxsRHw2M7MHG/gHuA+3s90WPwdON7PtgW8UaTOzk3Eb3AOBWZK2NLM/4HainwIulXRAO8JPkh4lG+wkSdrLIOA5M1suaVtghzr+7gO2kjREbvmhbKzjZtywAwCSPhCHjaxmnQUcLGlkXCNJYyRtXJG+JyPOMaU4hpvZfDM7BTcis1VYc3razM7FDZN8gCTp5WSDnSRJe7kRWE/SPHwuemaVJzN7GTeN+Hvgdtw85Itx+j8ijAWS7sOtTIFb0Hp/LBgbXRPeX3GrV/8j6QHc8tSOuBWmMifiw/K3AX8ruR8TC8vmAy/gFqx2B+ZJugfYF38pSJJeTWqJJ0nS5UgaGHPFwheMLTCzbBSTpBNkDztJku7giLAxfT++Avu8Hk5PkjQ92cNOkiRJkiYge9hJkiRJ0gRkg50kSZIkTUA22EmSJEnSBGSDnSRJkiRNQDbYSZIkSdIEZIOdJEmSJE3A/wOkMMET4zs9NQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11816f710>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_weights(clf_context, features_context, occupations)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### 2.4 Comparison of Model Weights\n",
    "\n",
    "Comparing the model coefficients side-by-side shows that the context model has more weights. Adding extra predictors can sometimes lead to overfitting. Even with L2 penalty, the weights for the binary context features end up having higher values than the other weights in their column so that they can override the effect of the industry features.\n",
    "\n",
    "Does this lead to overfitting? Certainly, the context model is more complex than the basic model. Using binary context features could lead to a higher variance model by enabling spurious associations or by making the model more sensitive to certain combinations of industries and job title words.\n",
    "\n",
    "However, the term \"overfitting\" may not be appropriate here because neither model is trained on data where the job title contains only the word `\"analyst\"`. Since those low-information job titles are used to evaluate the model, they still influence the model selection process. If we compare the models based on some out-of-sample validation set, we assume that future data will look like that data, both in content and distribution.\n",
    "\n",
    "In this hypothetical example, we expect that the word `\"analyst\"` should be treated differently based on the industry. Since the context model accomplishes that, it offers a richer representation for this task."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x108877940>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 2)\n",
    "show_weights(clf_basic, features_basic, occupations, ax=axes[0])\n",
    "show_weights(clf_context, features_context, occupations, ax=axes[1])\n",
    "axes[0].set_title(\"Basic Model Weights\")\n",
    "axes[1].set_title(\"Context Model Weights\")\n",
    "fig.set_size_inches((22, 6))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This hypothetical example shows a small training set and a small vocabulary. The `BinaryContextTransformer` can be useful if you need to efficiently generate contextual features for large datasets. A large dataset also offers a greater chance of generating unnecessary combinations. To determine if `BinaryContextTransformer` is appropriate for your task:\n",
    "\n",
    "- Check that there is a task or domain-related justification for binary context features\n",
    "- Use appropriate data cleaning and feature selection methods before applying this transformation"
   ]
  },
  {
   "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": "3.6.1"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}