accuracy/models-accuracy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Entities Recognition\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/entities](https://github.com/huseinzol05/Malaya/tree/master/session/entities)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from IPython.core.display import Image, display\n",
    "\n",
    "display(Image('ner-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Concat\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       1.00      1.00      1.00    498279\n",
    "       event       0.98      0.99      0.99      2217\n",
    "         law       0.99      0.99      0.99      1610\n",
    "    location       0.99      1.00      1.00     20194\n",
    "organization       0.99      0.99      0.99     26093\n",
    "      person       1.00      0.99      1.00     43377\n",
    "    quantity       1.00      1.00      1.00     13180\n",
    "        time       0.99      1.00      0.99     12750\n",
    "\n",
    " avg / total       1.00      1.00      1.00    617700\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       1.00      1.00      1.00    498587\n",
    "       event       0.98      0.99      0.98      2212\n",
    "         law       1.00      0.99      0.99      1746\n",
    "    location       0.99      1.00      1.00     20387\n",
    "organization       0.99      1.00      1.00     25376\n",
    "      person       1.00      1.00      1.00     43158\n",
    "    quantity       1.00      1.00      1.00     13581\n",
    "        time       0.99      1.00      0.99     12653\n",
    "\n",
    " avg / total       1.00      1.00      1.00    617700\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       1.00      1.00      1.00    497138\n",
    "       event       0.99      0.99      0.99      2331\n",
    "         law       0.99      0.99      0.99      1872\n",
    "    location       0.99      1.00      1.00     20671\n",
    "organization       0.99      1.00      0.99     25942\n",
    "      person       0.99      1.00      1.00     43511\n",
    "    quantity       1.00      1.00      1.00     13376\n",
    "        time       1.00      1.00      1.00     12859\n",
    "\n",
    " avg / total       1.00      1.00      1.00    617700\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Entity-Network\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       1.00      1.00      1.00    497198\n",
    "       event       0.98      0.95      0.96      2381\n",
    "         law       0.99      0.97      0.98      1881\n",
    "    location       0.99      0.99      0.99     20305\n",
    "organization       0.99      0.98      0.98     26036\n",
    "      person       0.99      0.99      0.99     43470\n",
    "    quantity       0.99      0.99      0.99     13608\n",
    "        time       0.98      0.99      0.98     12821\n",
    "\n",
    " avg / total       1.00      1.00      1.00    617700\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### CRF\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    quantity      0.991     0.991     0.991     13891\n",
    "    location      0.987     0.989     0.988     20798\n",
    "        time      0.987     0.977     0.982     13264\n",
    "      person      0.993     0.987     0.990     43590\n",
    "organization      0.974     0.973     0.973     25426\n",
    "       event      0.995     0.983     0.989      2417\n",
    "         law      0.994     0.988     0.991      1686\n",
    "\n",
    " avg / total      0.987     0.983     0.985    121072\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Attention\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       1.00      1.00      1.00    497073\n",
    "       event       0.99      0.97      0.98      2426\n",
    "         law       1.00      0.99      0.99      1806\n",
    "    location       1.00      1.00      1.00     20176\n",
    "organization       1.00      1.00      1.00     26044\n",
    "      person       1.00      1.00      1.00     44346\n",
    "    quantity       1.00      1.00      1.00     13155\n",
    "        time       0.99      1.00      1.00     12674\n",
    "\n",
    " avg / total       1.00      1.00      1.00    617700\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Language Detection\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/language-detection](https://github.com/huseinzol05/Malaya/tree/master/session/language-detection)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('language-detection-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### XGB\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       0.98      0.99      0.99      9424\n",
    "         eng       1.00      0.99      0.99      9972\n",
    "         ind       1.00      0.99      0.99     11511\n",
    "         zlm       1.00      1.00      1.00     10679\n",
    "\n",
    "   micro avg       0.99      0.99      0.99     41586\n",
    "   macro avg       0.99      0.99      0.99     41586\n",
    "weighted avg       0.99      0.99      0.99     41586\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Multinomial\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       1.00      0.97      0.99      9424\n",
    "         eng       0.99      1.00      0.99      9972\n",
    "         ind       1.00      1.00      1.00     11511\n",
    "         zlm       0.99      1.00      0.99     10679\n",
    "\n",
    "   micro avg       0.99      0.99      0.99     41586\n",
    "   macro avg       0.99      0.99      0.99     41586\n",
    "weighted avg       0.99      0.99      0.99     41586\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### SGD\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       OTHER       0.97      0.99      0.98      9424\n",
    "         eng       0.99      0.99      0.99      9972\n",
    "         ind       1.00      0.99      0.99     11511\n",
    "         zlm       1.00      1.00      1.00     10679\n",
    "\n",
    "   micro avg       0.99      0.99      0.99     41586\n",
    "   macro avg       0.99      0.99      0.99     41586\n",
    "weighted avg       0.99      0.99      0.99     41586\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Deep learning\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "       other       1.00      0.99      0.99      9445\n",
    "     english       1.00      1.00      1.00      9987\n",
    "  indonesian       1.00      1.00      1.00     11518\n",
    "       malay       1.00      1.00      1.00     10636\n",
    "\n",
    "   micro avg       1.00      1.00      1.00     41586\n",
    "   macro avg       1.00      1.00      1.00     41586\n",
    "weighted avg       1.00      1.00      1.00     41586\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## POS Recognition\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/pos](https://github.com/huseinzol05/Malaya/tree/master/session/pos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('pos-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Concat\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "        ADJ       0.99      1.00      1.00     22663\n",
    "        ADP       1.00      1.00      1.00     60199\n",
    "        ADV       1.00      1.00      1.00     23633\n",
    "        AUX       1.00      1.00      1.00      5249\n",
    "      CCONJ       1.00      1.00      1.00     18485\n",
    "        DET       1.00      1.00      1.00     19849\n",
    "       NOUN       1.00      1.00      1.00    135031\n",
    "        NUM       1.00      1.00      1.00     21842\n",
    "       PART       1.00      1.00      1.00      2900\n",
    "       PRON       1.00      1.00      1.00     23908\n",
    "      PROPN       1.00      1.00      1.00    113206\n",
    "      SCONJ       1.00      0.99      1.00      7304\n",
    "        SYM       1.00      1.00      1.00      1205\n",
    "       VERB       1.00      1.00      1.00     61222\n",
    "          X       0.97      0.99      0.98       154\n",
    "\n",
    "avg / total       1.00      1.00      1.00    516850\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "        ADJ       0.99      0.99      0.99     22879\n",
    "        ADP       1.00      1.00      1.00     60063\n",
    "        ADV       1.00      0.99      1.00     23653\n",
    "        AUX       1.00      1.00      1.00      5295\n",
    "      CCONJ       1.00      1.00      1.00     18395\n",
    "        DET       1.00      1.00      1.00     20088\n",
    "       NOUN       1.00      1.00      1.00    134736\n",
    "        NUM       1.00      1.00      1.00     21938\n",
    "       PART       0.99      1.00      0.99      3093\n",
    "       PRON       1.00      1.00      1.00     24060\n",
    "      PROPN       1.00      1.00      1.00    112859\n",
    "      SCONJ       0.99      0.99      0.99      7445\n",
    "        SYM       0.99      0.99      0.99      1219\n",
    "       VERB       1.00      1.00      1.00     60937\n",
    "          X       0.98      0.97      0.98       190\n",
    "\n",
    "avg / total       1.00      1.00      1.00    516850\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "        ADJ       0.99      0.99      0.99     22649\n",
    "        ADP       1.00      1.00      1.00     60088\n",
    "        ADV       0.99      1.00      0.99     24031\n",
    "        AUX       1.00      1.00      1.00      5279\n",
    "      CCONJ       1.00      1.00      1.00     18469\n",
    "        DET       1.00      1.00      1.00     20053\n",
    "       NOUN       1.00      1.00      1.00    134614\n",
    "        NUM       1.00      1.00      1.00     21703\n",
    "       PART       1.00      0.99      0.99      2956\n",
    "       PRON       1.00      1.00      1.00     23786\n",
    "      PROPN       1.00      1.00      1.00    113689\n",
    "      SCONJ       0.99      0.99      0.99      7315\n",
    "        SYM       1.00      1.00      1.00      1189\n",
    "       VERB       1.00      1.00      1.00     60827\n",
    "          X       0.97      0.99      0.98       202\n",
    "\n",
    "avg / total       1.00      1.00      1.00    516850\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### CRF\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "        PRON      0.998     0.997     0.998     47911\n",
    "         DET      0.990     0.993     0.991     39932\n",
    "        NOUN      0.988     0.988     0.988    270045\n",
    "        VERB      0.997     0.997     0.997    122015\n",
    "       PROPN      0.989     0.988     0.988    225893\n",
    "         ADP      0.997     0.998     0.997    120358\n",
    "         ADV      0.992     0.991     0.991     47753\n",
    "       CCONJ      0.997     0.998     0.997     36696\n",
    "         NUM      0.993     0.995     0.994     43748\n",
    "         ADJ      0.985     0.988     0.986     45244\n",
    "        PART      0.992     0.995     0.993      5975\n",
    "         AUX      1.000     1.000     1.000     10505\n",
    "       SCONJ      0.994     0.987     0.990     14798\n",
    "         SYM      0.998     0.997     0.998      2483\n",
    "\n",
    "   micro avg      0.992     0.992     0.992   1033356\n",
    "   macro avg      0.994     0.994     0.994   1033356\n",
    "weighted avg      0.992     0.992     0.992   1033356\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Entity-network\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "        ADJ       0.98      0.98      0.98     22626\n",
    "        ADP       0.99      0.99      0.99     60045\n",
    "        ADV       0.97      0.98      0.98     23537\n",
    "        AUX       0.99      0.99      0.99      5195\n",
    "      CCONJ       0.99      0.99      0.99     18357\n",
    "        DET       0.99      0.99      0.99     19762\n",
    "       NOUN       0.99      0.99      0.99    134505\n",
    "        NUM       0.99      0.99      0.99     22083\n",
    "       PART       0.97      0.97      0.97      2924\n",
    "       PRON       0.99      0.99      0.99     23783\n",
    "      PROPN       0.99      0.99      0.99    114144\n",
    "      SCONJ       0.96      0.95      0.95      7534\n",
    "        SYM       0.97      0.98      0.97      1335\n",
    "       VERB       0.99      0.99      0.99     60834\n",
    "          X       0.93      0.68      0.79       186\n",
    "\n",
    "avg / total       0.99      0.99      0.99    516850\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sentiment Analysis\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/sentiment](https://github.com/huseinzol05/Malaya/tree/master/session/sentiment)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('sentiment-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "   negative       0.79      0.81      0.80     70549\n",
    "   positive       0.79      0.77      0.78     64456\n",
    "\n",
    "avg / total       0.79      0.79      0.79    135005\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### BERT\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.75      0.83      0.79     70558\n",
    "    positive       0.79      0.70      0.75     64447\n",
    "\n",
    "   micro avg       0.77      0.77      0.77    135005\n",
    "   macro avg       0.77      0.77      0.77    135005\n",
    "weighted avg       0.77      0.77      0.77    135005\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bidirectional\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "   negative       0.82      0.79      0.80     70629\n",
    "   positive       0.77      0.81      0.79     64376\n",
    "\n",
    "avg / total       0.80      0.80      0.80    135005\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Entity-network\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "   negative       0.75      0.76      0.75     70301\n",
    "   positive       0.73      0.71      0.72     63700\n",
    "\n",
    "avg / total       0.74      0.74      0.74    134001\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.79      0.78      0.78     70568\n",
    "    positive       0.76      0.77      0.77     64437\n",
    "\n",
    "   micro avg       0.77      0.77      0.77    135005\n",
    "   macro avg       0.77      0.77      0.77    135005\n",
    "weighted avg       0.77      0.77      0.77    135005\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text-char\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "   negative       0.75      0.76      0.75     70301\n",
    "   positive       0.73      0.71      0.72     63700\n",
    "\n",
    "avg / total       0.74      0.74      0.74    134001\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hierarchical\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.81      0.80      0.80     70440\n",
    "    positive       0.78      0.79      0.79     64565\n",
    "\n",
    "   micro avg       0.80      0.80      0.80    135005\n",
    "   macro avg       0.80      0.80      0.80    135005\n",
    "weighted avg       0.80      0.80      0.80    135005\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "   negative       0.79      0.79      0.79     70613\n",
    "   positive       0.77      0.77      0.77     64392\n",
    "\n",
    "avg / total       0.78      0.78      0.78    135005\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Multinomial\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.78      0.84      0.81     70720\n",
    "    positive       0.80      0.74      0.77     64129\n",
    "\n",
    "   micro avg       0.79      0.79      0.79    134849\n",
    "   macro avg       0.79      0.79      0.79    134849\n",
    "weighted avg       0.79      0.79      0.79    134849\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### XGB\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.81      0.80      0.81     70356\n",
    "    positive       0.79      0.80      0.79     64493\n",
    "\n",
    "   micro avg       0.80      0.80      0.80    134849\n",
    "   macro avg       0.80      0.80      0.80    134849\n",
    "weighted avg       0.80      0.80      0.80    134849\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Toxicity Analysis\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/toxic](https://github.com/huseinzol05/Malaya/tree/master/session/toxic)\n",
    "\n",
    "Labels are,\n",
    "```python\n",
    "{0: 'toxic', 1: 'severe_toxic', 2: 'obscene', 3: 'threat', 4: 'insult', 5: 'identity_hate'}\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('toxic-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.75      0.58      0.65       797\n",
    "          1       0.46      0.36      0.41        77\n",
    "          2       0.71      0.59      0.64       423\n",
    "          3       0.00      0.00      0.00        30\n",
    "          4       0.68      0.56      0.61       384\n",
    "          5       0.40      0.23      0.29        74\n",
    "\n",
    "avg / total       0.69      0.54      0.60      1785\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Entity-network\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.77      0.61      0.68       787\n",
    "          1       1.00      0.08      0.14        92\n",
    "          2       0.75      0.66      0.71       432\n",
    "          3       0.44      0.19      0.27        21\n",
    "          4       0.68      0.61      0.65       400\n",
    "          5       0.57      0.05      0.09        79\n",
    "\n",
    "avg / total       0.75      0.57      0.62      1811\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.80      0.47      0.59       815\n",
    "          1       0.58      0.19      0.29        94\n",
    "          2       0.90      0.38      0.54       425\n",
    "          3       0.40      0.06      0.11        31\n",
    "          4       0.85      0.35      0.49       410\n",
    "          5       0.35      0.08      0.13        74\n",
    "\n",
    "avg / total       0.80      0.39      0.52      1849\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hierarchical\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.69      0.56      0.62       751\n",
    "          1       0.68      0.20      0.31        86\n",
    "          2       0.84      0.48      0.61       412\n",
    "          3       1.00      0.04      0.07        28\n",
    "          4       0.70      0.55      0.62       356\n",
    "          5       0.57      0.27      0.36        60\n",
    "\n",
    "avg / total       0.73      0.50      0.58      1693\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Logistic Regression\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.98      0.27      0.43       805\n",
    "          1       0.50      0.02      0.04        88\n",
    "          2       0.99      0.30      0.46       460\n",
    "          3       0.00      0.00      0.00        32\n",
    "          4       0.87      0.22      0.35       420\n",
    "          5       0.00      0.00      0.00        68\n",
    "\n",
    "avg / total       0.88      0.24      0.38      1873\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Multinomial\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.81      0.52      0.63       805\n",
    "          1       0.44      0.35      0.39        88\n",
    "          2       0.76      0.49      0.59       460\n",
    "          3       0.00      0.00      0.00        32\n",
    "          4       0.68      0.47      0.56       420\n",
    "          5       0.15      0.09      0.11        68\n",
    "\n",
    "avg / total       0.71      0.47      0.56      1873\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0       0.68      0.64      0.66       799\n",
    "          1       0.43      0.22      0.29        81\n",
    "          2       0.77      0.59      0.67       423\n",
    "          3       0.57      0.16      0.25        25\n",
    "          4       0.70      0.53      0.61       404\n",
    "          5       0.33      0.01      0.02        81\n",
    "\n",
    "avg / total       0.68      0.55      0.60      1813\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Subjectivity Analysis\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/subjectivity](https://github.com/huseinzol05/Malaya/tree/master/session/subjectivity)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('subjectivity-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.82      0.83      0.82       983\n",
    "    positive       0.83      0.83      0.83      1010\n",
    "\n",
    "   micro avg       0.83      0.83      0.83      1993\n",
    "   macro avg       0.83      0.83      0.83      1993\n",
    "weighted avg       0.83      0.83      0.83      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### BERT\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.75      0.95      0.84      1024\n",
    "    positive       0.93      0.67      0.78       969\n",
    "\n",
    "   micro avg       0.82      0.82      0.82      1993\n",
    "   macro avg       0.84      0.81      0.81      1993\n",
    "weighted avg       0.84      0.82      0.81      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bidirectional\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.82      0.89      0.85       989\n",
    "    positive       0.88      0.81      0.84      1004\n",
    "\n",
    "   micro avg       0.85      0.85      0.85      1993\n",
    "   macro avg       0.85      0.85      0.85      1993\n",
    "weighted avg       0.85      0.85      0.85      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Entity-network\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.89      0.87      0.88       984\n",
    "    positive       0.87      0.89      0.88      1009\n",
    "\n",
    "   micro avg       0.88      0.88      0.88      1993\n",
    "   macro avg       0.88      0.88      0.88      1993\n",
    "weighted avg       0.88      0.88      0.88      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.88      0.91      0.89      1003\n",
    "    positive       0.90      0.87      0.89       990\n",
    "\n",
    "   micro avg       0.89      0.89      0.89      1993\n",
    "   macro avg       0.89      0.89      0.89      1993\n",
    "weighted avg       0.89      0.89      0.89      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text-char\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "   negative       0.88      0.88      0.88      1002\n",
    "   positive       0.88      0.87      0.88       991\n",
    "\n",
    "avg / total       0.88      0.88      0.88      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hierarchical\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.83      0.87      0.85      1017\n",
    "    positive       0.85      0.81      0.83       976\n",
    "\n",
    "   micro avg       0.84      0.84      0.84      1993\n",
    "   macro avg       0.84      0.84      0.84      1993\n",
    "weighted avg       0.84      0.84      0.84      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.85      0.78      0.81       987\n",
    "    positive       0.80      0.86      0.83      1006\n",
    "\n",
    "   micro avg       0.82      0.82      0.82      1993\n",
    "   macro avg       0.82      0.82      0.82      1993\n",
    "weighted avg       0.82      0.82      0.82      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Multinomial\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.91      0.85      0.88       999\n",
    "    positive       0.86      0.92      0.89       994\n",
    "\n",
    "   micro avg       0.89      0.89      0.89      1993\n",
    "   macro avg       0.89      0.89      0.89      1993\n",
    "weighted avg       0.89      0.89      0.89      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### XGB\n",
    "\n",
    "```text\n",
    "\n",
    "              precision    recall  f1-score   support\n",
    "\n",
    "    negative       0.86      0.85      0.85      1003\n",
    "    positive       0.85      0.86      0.85       990\n",
    "\n",
    "   micro avg       0.85      0.85      0.85      1993\n",
    "   macro avg       0.85      0.85      0.85      1993\n",
    "weighted avg       0.85      0.85      0.85      1993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Emotion Analysis\n",
    "\n",
    "Trained on 80% of dataset, tested on 20% of dataset. All training sessions stored in [session/emotion](https://github.com/huseinzol05/Malaya/tree/master/session/emotion)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('emotion-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.80      0.80      0.80      3827\n",
    "       fear       0.77      0.78      0.78      3760\n",
    "        joy       0.81      0.78      0.80      3958\n",
    "       love       0.82      0.86      0.84      3099\n",
    "    sadness       0.73      0.76      0.74      3119\n",
    "   surprise       0.79      0.74      0.77      1940\n",
    "\n",
    "avg / total       0.79      0.79      0.79     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### BERT\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.73      0.83      0.78      3747\n",
    "       fear       0.70      0.84      0.77      3789\n",
    "        joy       0.74      0.80      0.77      3929\n",
    "       love       0.82      0.76      0.79      3081\n",
    "    sadness       0.82      0.60      0.69      3168\n",
    "   surprise       0.85      0.63      0.72      1989\n",
    "\n",
    "avg / total       0.77      0.76      0.76     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bidirectional\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.81      0.80      0.81      3726\n",
    "       fear       0.77      0.78      0.77      3806\n",
    "        joy       0.83      0.81      0.82      3975\n",
    "       love       0.86      0.83      0.85      2992\n",
    "    sadness       0.75      0.78      0.77      3293\n",
    "   surprise       0.77      0.79      0.78      1911\n",
    "\n",
    "avg / total       0.80      0.80      0.80     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Entity-network\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.82      0.72      0.77      3717\n",
    "       fear       0.72      0.77      0.75      3743\n",
    "        joy       0.77      0.74      0.76      4050\n",
    "       love       0.81      0.81      0.81      2992\n",
    "    sadness       0.71      0.74      0.72      3274\n",
    "   surprise       0.72      0.80      0.76      1927\n",
    "\n",
    "avg / total       0.76      0.76      0.76     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.82      0.75      0.78      3754\n",
    "       fear       0.71      0.81      0.75      3837\n",
    "        joy       0.76      0.79      0.78      3844\n",
    "       love       0.83      0.83      0.83      3065\n",
    "    sadness       0.75      0.75      0.75      3241\n",
    "   surprise       0.79      0.64      0.71      1962\n",
    "\n",
    "avg / total       0.77      0.77      0.77     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fast-text-char\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.79      0.75      0.77      3803\n",
    "       fear       0.73      0.73      0.73      3784\n",
    "        joy       0.71      0.77      0.74      3872\n",
    "       love       0.81      0.80      0.80      3052\n",
    "    sadness       0.72      0.70      0.71      3205\n",
    "   surprise       0.73      0.70      0.72      1987\n",
    "\n",
    "avg / total       0.75      0.74      0.75     19703\n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hierarchical\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.81      0.79      0.80      3786\n",
    "       fear       0.78      0.79      0.78      3754\n",
    "        joy       0.81      0.82      0.82      3886\n",
    "       love       0.85      0.84      0.85      3022\n",
    "    sadness       0.76      0.80      0.78      3300\n",
    "   surprise       0.81      0.75      0.78      1955\n",
    "\n",
    "avg / total       0.80      0.80      0.80     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.80      0.79      0.80      3774\n",
    "       fear       0.78      0.75      0.77      3759\n",
    "        joy       0.79      0.80      0.79      3944\n",
    "       love       0.83      0.84      0.84      3033\n",
    "    sadness       0.75      0.75      0.75      3272\n",
    "   surprise       0.76      0.80      0.78      1921\n",
    "\n",
    "avg / total       0.79      0.79      0.79     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Multinomial\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.72      0.82      0.77      3833\n",
    "       fear       0.68      0.80      0.74      3802\n",
    "        joy       0.68      0.84      0.75      3924\n",
    "       love       0.85      0.71      0.78      2981\n",
    "    sadness       0.81      0.67      0.73      3189\n",
    "   surprise       0.80      0.36      0.50      1974\n",
    "\n",
    "avg / total       0.75      0.73      0.73     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### XGB\n",
    "\n",
    "```text\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "      anger       0.80      0.80      0.80      3769\n",
    "       fear       0.79      0.76      0.78      3808\n",
    "        joy       0.79      0.81      0.80      3913\n",
    "       love       0.84      0.85      0.84      2998\n",
    "    sadness       0.76      0.75      0.76      3250\n",
    "   surprise       0.77      0.77      0.77      1965\n",
    "\n",
    "avg / total       0.79      0.79      0.79     19703\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dependency parsing\n",
    "\n",
    "Trained on 90% of dataset, tested on 10% of dataset. All training sessions stored in [session/dependency](https://github.com/huseinzol05/Malaya/tree/master/session/dependency)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAukAAAHbCAYAAAB2ngQwAAAYTmlDQ1BJQ0MgUHJvZmlsZQAAWIWVWQk4lN3bP8/sM/YZ+77vZN/3fd93EsY2thhrKJFkK1EhSiVZK5VCJSItSll6SZJIlkqhorLkeyz1vv/3f13fd31nrvM8v7nPfe7trPcMAFz7fSMjwxCMAIRHxFAdTA343dw9+LETAII/LEATIHzJ0ZH6dnZWAC6/3/9ZlgZhXrg8l9mQ9d/t/2th8g+IJgMA2cHYzz+aHA7j6wCg0smR1BgAMGowXSg+JnIDe8GYmQobCOPIDRy0hdM3sN8WLt7kcXIwhPFFAHC0vr7UIADom2E6fxw5CJZDPwS3ESP8KREw6yyMdcjBvv4AcEnDPNLh4bs3sBuMxf3+ISfoP2T6/ZHp6xv0B2/5sllwRpToyDDfPf/PcPzfJTws9rcOUbjSBlPNHDZ8huM2FLrbcgPTwng2ws/GFsZEGP+g+G/ywxhBCI41c97iR3CTow3hmAFWGMv5+xpZwpgbxiYRYTZW23S/QIqJOYzhGYJIoMSYO233zQyINnbclnmKutvB9jcOpBrqb/e97Evd1LvB3xkb6qy/LX8oOMD8t/xvicFOrls2IwlxFBcbGNPDmDU61NFyiwcpnBhsaPObhxrrsGG/MIw1AiJMDbbkI70DqSYO2/zU8Ojf/iIzgynmNtu4JCbYyWxbzkWy76b97DBuDojQd/4tJyDazeq3L/4BRsZbviN7AyKct/1FjkXGGDhs9/0SGWa3zY8iBISZbtAFYcwdHee43RelEwNPyC35KJvIGDunLTtRfiG+FnZb9qASgBUwBEaAH8TC1Q/sBiGA8my2aRb+ttViAnwBFQSBACCzTfndw3WzJQJ+OoJE8AlGASD6Tz+DzdYAEAfT1/5Qt54yIHCzNW6zRyiYgnE4sARh8PfYzV4Rf7S5gHcwhfJf2smwrWFw3Wj7b5o+TLHapsT+lsvP8JsTY4wxwphhTDASKE6UDkoTZQU/9eCqgFJDqf+29m9+9BS6D/0WPYAeQ7/cRUmj/ssffmANxmANJts++/3TZ5QoLFUZZYDShuXDslGsKE4gg1KCNemjdGHdyjDVcNvyDe//Lfs/fPhH1Lf58HJ4BJ4Nr4cX/3dPekl65T9SNmL6zwht2er3J66Gf1r+rd/wH5H2h9+W/+ZEZiIbkA+R7cguZAuyCfAj25DNyG7knQ38Zxa925xFv7U5bNoTCsuh/Jc+322dG5GMlquTey+3utUWE5AQs7HADHdH7qFSgoJj+PXhnT+A3zyCLCvNryAnrw7AxjmytU19ddg8HyDWnr9p5IMAqM4DgF/+mxb+FYArBHgbtf6bJuINLzMMANVT5Fhq3BYNtfFAAwJggFcUB+AFQkAc9kcBqMDnlR4wBhbAFjgBd+ANRzkYns9UEA+SQSrIADngKDgBSsAZcB5Ug0vgGmgCLaAdPABPQC8YAK/g2TMJPoJ5sARWIAjCQnQQCeKA+CARSApSgNQgHcgYsoIcIHfIBwqCIqBYKBk6AOVABVAJdA6qga5CN6F2qAvqg15C49B76Au0jEAiaBHMCB6EKGIHQg2hj7BEOCF2IoIQUYhERDriCKIYUY64iGhEtCOeIAYQY4iPiEUkQNIgWZECSBmkGtIQaYv0QAYiqch9yGxkIbIceRl5Cx7n58gx5CzyJwqDIqH4UTLwDDZDOaPIqCjUPlQuqgRVjWpEdaKeo8ZR86hfaDo0N1oKrYE2R7uhg9Dx6Ax0IboSfQN9H15Nk+glDAbDihHDqMKr0R0TgknC5GJOY+oxdzF9mAnMIhaL5cBKYbWxtlhfbAw2A3sSexHbhu3HTmJ/4GhwfDgFnAnOAxeBS8MV4mpxrbh+3DRuBc+IF8Fr4G3x/vg9+Dx8Bf4Wvgc/iV8hMBHECNoEJ0IIIZVQTLhMuE8YIXyloaERpFGnsaeh0OynKaa5QvOIZpzmJy2RVpLWkNaLNpb2CG0V7V3al7Rf6ejoROn06DzoYuiO0NXQ3aMbpftBT6KXpTen96dPoS+lb6Tvp//MgGcQYdBn8GZIZChkaGDoYZhlxDOKMhoy+jLuYyxlvMn4gnGRicQkz2TLFM6Uy1TL1MU0Q8QSRYnGRH9iOvE88R5xgoQkCZEMSWTSAVIF6T5pkhnDLMZszhzCnMN8ifkZ8zwLkUWJxYUlgaWU5Q7LGCuSVZTVnDWMNY/1Gusg6zIbD5s+WwBbFttltn627+xc7HrsAezZ7PXsA+zLHPwcxhyhHPkcTRyvOVGckpz2nPGcZZz3OWe5mLk0uchc2VzXuIa5EdyS3A7cSdznubu5F3l4eUx5InlO8tzjmeVl5dXjDeE9ztvK+56PxKfDR+E7ztfG94GfhV+fP4y/mL+Tf16AW8BMIFbgnMAzgRVBMUFnwTTBesHXQgQhNaFAoeNCHULzwnzC1sLJwnXCwyJ4ETWRYJEikYci30XFRF1FD4k2ic6IsYuZiyWK1YmNiNOJ64pHiZeL/yWBkVCTCJU4LdEriZBUlgyWLJXskUJIqUhRpE5L9UmjpdWlI6TLpV/I0Mroy8TJ1MmMy7LKWsmmyTbJft4hvMNjR/6Ohzt+ySnLhclVyL2SJ8pbyKfJ35L/oiCpQFYoVfhLkU7RRDFFsVlxQUlKKUCpTGlImaRsrXxIuUN5TUVVhapyWeW9qrCqj+op1RdqzGp2arlqj9TR6gbqKeot6j81VDRiNK5pzGnKaIZq1mrOaIlpBWhVaE1oC2r7ap/THtPh1/HROaszpiug66tbrvtWT0jPX69Sb1pfQj9E/6L+ZwM5A6rBDYPvhhqGew3vGiGNTI2yjZ4ZE42djUuMR00ETYJM6kzmTZVNk0zvmqHNLM3yzV6Y85iTzWvM5y1ULfZadFrSWjpalli+tZK0olrdskZYW1gfsx6xEbGJsGmyBbbmtsdsX9uJ2UXZ3bbH2NvZl9pPOcg7JDs8dCQ57nKsdVxyMnDKc3rlLO4c69zhwuDi5VLj8t3VyLXAdcxth9tetyfunO4U92YPrIeLR6XHoqex5wnPSS9lrwyvwZ1iOxN2dnlzeod539nFsMt3V4MP2sfVp9Zn1dfWt9x30c/c75TfPNmQXET+6K/nf9z/fYB2QEHAdKB2YEHgTJB20LGg98G6wYXBsxRDSgllIcQs5EzI91Db0KrQ9TDXsPpwXLhP+M0IYkRoROdu3t0Ju/sipSIzIseiNKJORM1TLamV0VD0zujmGGb4wt4dKx57MHY8TieuNO5HvEt8QwJTQkRC9x7JPVl7phNNEi8koZLISR3JAsmpyeN79fee2wft89vXkSKUkp4yud90f3UqITU09WmaXFpB2rcDrgdupfOk70+fOGh6sC6DPoOa8eKQ5qEzmahMSuazLMWsk1m/sv2zH+fI5RTmrOaScx8flj9cfHj9SOCRZ3kqeWVHMUcjjg7m6+ZXFzAVJBZMHLM+1nic/3j28W8ndp3oKlQqPFNEKIotGiu2Km4+KXzy6MnVkuCSgVKD0vpT3KeyTn0/7X+6v0yv7PIZnjM5Z5bPUs4OnTM911guWl54HnM+7vxUhUvFwwtqF2oqOStzKteqIqrGqh2qO2tUa2pquWvz6hB1sXXvL3pd7L1kdKn5sszlc/Ws9TlXwJXYKx+u+lwdvGZ5raNBreHydZHrp26QbmQ3Qo17GuebgpvGmt2b+25a3Oy4pXnrxm3Z21UtAi2ld1ju5LUSWtNb19sS2xbvRt6dbQ9qn+jY1fHqntu9vzrtO5/dt7z/6IHJg3sP9R+2PdJ+1NKl0XXzsdrjpicqTxq7lbtvPFV+euOZyrPGHtWe5l713lt9Wn2t/br97c+Nnj/4y/yvJwM2A32DzoNDL7xejA35D828DHu5MBw3vPJq/wh6JPs14+vCUe7R8jcSb+rHVMbujBuNd791fPtqgjzx8V30u9XJ9Cm6qcJpvumaGYWZlvcm73s/eH6Y/Bj5cWU24xPTp1OfxT9fn9Ob6553m59coC6sf8n9yvG16pvSt45Fu8XRpfClle/ZPzh+VP9U+/lw2XV5eiV+FbtavCaxduuX5a+R9fD19Uhfqu/mVQAJV0RgIABfqgCgcweA1AsAwXMrz9suSPjygYDfLpAs9BGRDp+oPagMtAkGiXmCLcZF4K0IEjRYmlnafrom+iqGSsZ6pmZiB+kJcy/LEOsbthn2jxwLnMtcazwIXiwfgZ9OgChIFGIVZhdhE2UX4xbnkeCX5JcSlBaWEZUV2yEtJyevqKCiqKGkq2ysYq5qrmaibqJhommopa+tpaOhq6Qnqy9qwGPIbEQwWjf+ajJl+tKs27zFotrymFWKdYiNm62xnbK9mAOXI6MTzhnpArki3FDueA9GTw4v4Z0y3hK7hH34fDn9WMgkf2IAKZA1iCtYkCIdohpqEuYSTolI3l0QWRF1llocnR+TG5sVlx1/JKF4T3Via9KrvWCfdMqu/SdTXx0QTN99sP0QJlMoSyHbIMcxN/Bw4pH8vOqjd/OHCxaPM52QKbQoCiw+cLKs5GZp/6l3pxfPYM9ynJMs1zpvW+F3IabyYFVhdXXNzdrHdcMXP1z6WY+7wnZV/Jpug/v1qBtZjaeb6pvbbnbd6rnd2/LkTkfr1bbSuyntuzo07hHvTXXevF/74NTDnEcJXX6PzZ/IdtN3zz69/+xUT2SvQR+pb6L/2vPUv+wHRAZRg+9fdA/VvywYjnnlMqL2mvP16ujom/axC+NZb3dPOL/TmhSGZ9nS9F8z198XfUj5GDZL/kT+HDmXM39jYe6r3rdzS6TvxT+llp+tpvzSWF//x/grIGdQBWhLDAvmNbYBl4sPIhjRSNIy0K7STdMPMQwxvmF6R/xE+sq8xLLGusK2xv6LY41ziesr9xzPFO8IXz//fYGbgpVCOcJhIlaikmJ4sQ/iXRI1ktlSFGlLGRlZOtm5HX1y1+WLFJIVyUr2ygYqCqoCakS1dfXPGiOaXVqN2uU6ubrxej76FgYKhpxGCKP3xs9Mrpjmm0WbO1moWLJZrli9sb5nU2ubb5dkH+jg6KjvJO8s4EJyxbouu310H/Ho9rzjVb/zrPexXYd8kn2pfhSyr79HgFOgfZBNsCXFMsQsVDNMNlwggmU3TSQicjXqB/Vn9FosOo4YL5SgsccpMTqpMLll71QKzX6+VJk07QM26X4H4zMOH6rMbMsazv6ey3xY4Yh9XsTRw/l1BY+OvTu+XshZpFxsdzK05GDpmVPNp3vLZs78OsdcLnFeu8LuArkytupQdTG8z3XXzV0iXlasd7wSdTXvWl1D5/WRG1+aMM0cNyVvady2aHG7E9ga05ZyN7X9QMfBexmdh+5nPsh+mPvocNfhx4efHO7OfZrzLKvnUG96X2r/3udxf0UN7B6MfBEzlPTy4PCxV+UjDa8fjL5882kcvCVOCL6Tn9SZMp/2mzn7/tNH5dmkT62ff81rLsR9ufz13SL7kuX3lB8NP6dXuFcd1rJ/dW6PvzFCH7kD+RnVjj6EccSKYxdwN/EZBAcabppR2vN04fTqDAiGdsZ0JgsiA7GXdJTZloWB5SlrNpsJO8TezBHBKcQ5xJXDrcP9iaeU14z3G18Zvxn/Z4HjghqCI0J7hfmFW0W8RVZFi8WUxLrFA8RXJY5JSkm2STlKTUmnyojIDMnm7jDY8U2uSt5TgU6hTTFSSUCpXzlNRUFlXDVPTVvtk3qphrnGouZ5LXutX9p1Ou66WN0bemR9ov5dg0hDfsNeozRjJeNpkxJTW/jecds8ykLK4p1lmZWHNav1c5sCWwc7kt2g/UkHb0dhxw9OV50TXYxdGVyH3Srdoz0MPGk9B73O7Az2VvBe2XXfJ9/Xy0/Cb4nc6X8swDdQMQgVNBhcS0kJcQqVDkOHvQm/FVG8Oz7SNUqDyhuNip6NGYhtj6uPL0vI25OaGJ8Umuy/d+c+txSn/Q6p9mn2BxzSnQ66Z+w8FJAZmhWdnZKTmVtwuOxITV7j0Xv5fQWjxz6fQBVKFHkVHz15v2TllOxpv7ITZx6fXS1XOB9QUXKhpwpVrVUTX1tf9/GS5OWQ+torc9dUGvZf727kaApr7rzFdzul5W2rVVtLu3zHxU6p+1cfGjwafpzQzfe0t+dwn9Nz0QEw+HHo3fCH1+CNyPiuidop9EziR/CpYp78VXdJ7afzavHG+G/93rdRMCoAnDgEwMbvOQ41AOReBEBsDwBscO5pRweAkzpACJgCaLEDQBbaf84PCE48CYAEeIAkUANmcH4ZBueURaAedIEJsAaxQ8qQIxQNnYCaoddwzieNcEGkIuoRo0gGpAEyHnkJOQVnaV6oUtQrOBPzQV9Af8KoYFIxz7A82HBsG46Eo+Da8Zz4WHw/QZFQRFilIdM8pVWnraZjp8ulR9An0X9niGVYYkxkgpiyiazECpIaqZc5hAXLcoHVmHWKLZNdir2XI4aTg7OVy5+bhvsajwcvkvcSnyecEfQJ5AnaCjEJPRcuEvEUFRCdErsoHi2hIQlJdknlS3vCs3Netn9Hi1ylfIHCPkWKkqOyhgqfKqQ6ptaiflwjRFNbi15rRLtGJ0ZXVw+n16ffYHDdsMnolnGryT3TLrMe80GLUctpqwXrFVucHau9qIOao5UT2TnZpdi11W3Gg+Sp7xW5s9x7wIfgq++XRG72/x6oFpQU3B5CCHUOqwhf3G0WWRY1F60VkxM7Gq+UcHTPQpJr8oN92imtqZZpE+lZGdqZIKsv58rhU3kF+WbHkMfvF+YXB5QYnpIuEzwrUq5UYVMZVV1a++QSqFe9atPgfiO4KfnmidvX7vS3LXXwdpo9iHl09vGz7rUemb6dz48M3B0iDZNHLo3OjnNPqE3qTcu/p//wYvbI5x1z7QtmXzq/KSyWLC3/sP95YXlhVWMt5dfdzf1ja/yJ8PhLAFVgAlxBCNgHjoM60AlGwQ+IBMlBNlAEdBRqgF4iAEICzvLTEFcRb+E83gqZjmxDrqC0UQdQ3Wh2dCC6EYPHeGMasYzYMOwTnDQuD7eI98I/IMgSimiQNFE047TOtI/pDOla6bXo78BZ7CNGe8ZROE9dJx4nyZKeMkfAmWczqy8bDVszeyAHK8dDzj1c0lzj3EU8trw43g6+/fwGAhiBp4KFQr7CssKrIt2iZWJR4sYSXBJfJB9LnZdOkfGU1dwhIccuj5dfVZhTnFB6ofxY5bbqRbUS9UMaVE1PLUNtSR1GnUXdYb1W/XqDq4YNRk3Gt03aTDvNHpv3WrywfGM1bb1gs2KHs2d1EHNUd7J29nfZ61ridtN92GPNS3CnhXfMrrM+PX4QWcU/IqAmcCpYlBIScjV0Odw0onD3TJQWdW90Wywqziq+KGEqUT3pSPL0PuOU6lT6tD0HpuH9pDfTIuthjllu9xGHvLH8lGO8x+8WBhbTn2wu9T9NKntwdm+5yvkvF65WxdZo1WEuDly+cCX5mtd1lUb6pomb128fuGPTxn53vKOmk/pA6xG2a/BJzdP9PV59Os9FBpgGHw05v5x8lfiaefTamNP46kT1pPs0w0zXh8xZy8+Mcy8Wzn4NWVT5jvjRs1y6GvRLcXv8kQADaDd3AHGgAs8ANxAODoIz4DYYhte/IGQBxUIV0CCCBmEEr/wOJA5pjzyD/IKyQFWh8Wgq+g3GCV7tNtgBHBn3E19IUCdM0pyk1aMdoUui56fvYohnlGScYDpD9CNJkL4zP2QpY01i82TX45DiZOei4UZwr/Is867yAwEsfAPlEZYV0RZ1EAsS3y9xUvIGnHfPyzLuUJBzld+nUKHYo7SiIqHqrlag3q/JrOWuXaEzp6etf9jgjZGicY7JuJmWeaHFFys760u2tHZh9o8dJZ1ynD+4WrjVeuA9KV4PvUV3HfSZ9DMkVwYgA/2D7lFEQzJCZ8KtIuojWaISqGMxRrGX49kT9u35mOQGr1OVlKpUjrQj6aiDyRlfMj2yrmav5zodrjqyfNQx//IxwnHKiQdFUsW5J+dKXU/dKRM9kw/v/f7nuy9oVlZVM9Uk1k5ddLzUUi96Je/qUoP39QeNMk1Hm+dv2d++fIfQGtjW2k7sCLjXeB/1wO5h6aOJxxJPKN2VT8d7OHvt+w72X3/+doAwKPfCYYj68shwzat7IwOvp0YX3qyOQ2+xE5h3mEkwuTz1aXp05un75g/lHzNnIz5Zf5aaw869mW9eyPri8VXi65dvLYtpS0bfMd87f6T81Py5sHxhxWOVsNq4Rv5F9+vauvvG+EcHKipsHh8QrQEA6NH19a+iAGALAFjLX19fKV9fXzsPJxsjANwN2/oPafOsYQTgLNcG6rsy+frf/+X8D78U1VXQOwEbAAABnWlUWHRYTUw6Y29tLmFkb2JlLnhtcAAAAAAAPHg6eG1wbWV0YSB4bWxuczp4PSJhZG9iZTpuczptZXRhLyIgeDp4bXB0az0iWE1QIENvcmUgNS40LjAiPgogICA8cmRmOlJERiB4bWxuczpyZGY9Imh0dHA6Ly93d3cudzMub3JnLzE5OTkvMDIvMjItcmRmLXN5bnRheC1ucyMiPgogICAgICA8cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0iIgogICAgICAgICAgICB4bWxuczpleGlmPSJodHRwOi8vbnMuYWRvYmUuY29tL2V4aWYvMS4wLyI+CiAgICAgICAgIDxleGlmOlBpeGVsWERpbWVuc2lvbj43NDU8L2V4aWY6UGl4ZWxYRGltZW5zaW9uPgogICAgICAgICA8ZXhpZjpQaXhlbFlEaW1lbnNpb24+NDc1PC9leGlmOlBpeGVsWURpbWVuc2lvbj4KICAgICAgPC9yZGY6RGVzY3JpcHRpb24+CiAgIDwvcmRmOlJERj4KPC94OnhtcG1ldGE+CmRCZugAAEAASURBVHgB7N0JfE1n+sDxJxuyCUJJKyhSaq9paVWtRadKlaq2uhlDtUaX/2h1mU7RVhltFWOmC1PKzLQd1DKmRct0qKltij/+tURCLBFCNgmy/c/z6r2TCHLrnhw3N7/z+UTO8p73nPM9181z3/u87wnIy8srFCYEEEAAAQQQQAABBBDwGYFAnzkTTgQBBBBAAAEEEEAAAQSMAEE6LwQEEEAAAQQQQAABBHxMgCDdx24Ip4MAAggggAACCCCAAEE6rwEEEEAAAQQQQAABBHxMgCDdx24Ip4MAAggggAACCCCAAEE6rwEEEEAAAQQQQAABBHxMgCDdx24Ip4MAAggggAACCCCAAEE6rwEEEEAAAQQQQAABBHxMgCDdx24Ip4MAAggggAACCCCAAEE6rwEEEEAAAQQQQAABBHxMgCDdx24Ip4MAAggggAACCCCAAEE6rwEEEEAAAQQQQAABBHxMgCDdx24Ip4MAAggggAACCCCAAEE6rwEEEEAAAQQQQAABBHxMINiu89mwYYMsX75ccnNzpVu3btKlS5cLVr1p0yZZtGiRnDlzRtq3by/9+/eXwMBAycnJkblz58qOHTukfv360rdvX2ncuLGpIy8vz+zz7bffSlRUlNx///3StGnTC9bPSgQQQAABBBBAAAEEyruALS3pO3fulFGjRkloaKh06NBBxo0bJ6tXry5hc/z4cRk5cqTUq1dPunfvLvPnz5eZM2eacjNmzJC1a9dK7969JSQkRIYNGybHjh0z2z766CP5y1/+Ij179pRrrrlGHn30UcnKyipRPysQQAABBBBAAAEEEPAHAVta0hcuXCj9+vWT0aNHG5Ndu3aZALxr167FjFatWiWtW7eW4cOHm/WFhYUydepUGTJkiCxbtkwmTZok7dq1kx49esiKFStk48aNJpjXQP7jjz+WJk2amP3q1q0r2dnZEhERUax+FhBAAAEEEEAAAQQQ8AcBW1rSDxw4IM2aNXN7xMbGSnx8vHvZNXPq1CkJDw93LUp+fr6cPHlSMjIyTNAdFhZmtmnwrmkzCQkJkpycbNb98MMP8vjjj8tbb70lcXFxctVVV7nrYQYBBBBAAAEEEEAAAX8SsC1IL9qqXbt2bRN8nw+lLevr1q2TJUuWyObNm2XevHmmiAbrvXr1klmzZsnWrVtNbroG75rS4grS//rXv5pcdw3cNRXm7Nmz51dfYtmVSlNiAysQQAABBBBAAAEEEPBhAVvSXbQFPC0tzX2Zp0+fdqemuFdaMw0aNDCpLlOmTDGrNWjXoLtmzZoydOhQGT9+vNmurfItWrQw5bXTqE4vvviiSZUZOHCgSYFZv3693HbbbWab5rprLnvRSevU9TfeeGPR1cwjgAACCCCAAAIIIFAmAs2bNzd9K+2o3JYgPSYmRo4cOeI+n6SkJJOS4l7x44yW6dy5swnINaVFO5dqkB4QECCpqakybdo00YBflzUY17SZ6Ohos7e2zuukI8FoAJ6ZmWmW9R9d1qD9/ElHj2nTps35q1lGAAEEEEAAAQQQQMCnBWxJd9EW8QULFkhKSops377dzLdq1cpc+LZt20wnUF1ITEyUwYMHy4kTJ0wrtw65qJ1ENSifPn26zJ49WzR4X7x4sQnCW7ZsKQ0bNjT555988okZpnHNmjWiOfAdO3b0aVhODgEEEEAAAQQQQACByxWwpSVdR3bRluw+ffqY89DhFV3zul4Dax0+UUdu6dSpkwwYMMB0FNX5QYMGmX003UVHd1m6dKkZQ33ixIkSGRlpto0dO9aMHKN56TrpMI5Vq1Y18/yDAAIIIIAAAggggIC/CQRYOd+Fdl2Ujmuu6SiuFJWL1Zuenm5Gb9E0laJTQUGBaKqMDrEYFBRUdJMpf+jQIalVq1axEWKKFTpvQdNdLpQGc14xFhFAAAEEEEAAAQQQ8CkBW1rSXVekAbQnkz419EKTBvj6tNELTfqAowZWx1MmBBBAAAEEEEAAAQT8XcCWnHR/R+L6EEAAAQQQQAABBBBwUoAg3UltjoUAAggggAACCCCAgAcCBOkeIFEEAQQQQAABBBBAAAEnBQjSndTmWAgggAACCCCAAAIIeCBAkO4BEkUQQAABBBBAAAEEEHBSgCDdSW2OhQACCCCAAAIIIICABwIE6R4gUQQBBBBAAAEEEEAAAScFCNKd1OZYCCCAAAIIIIAAAgh4IECQ7gESRRBAAAEEEEAAAQQQcFKAIN1JbY6FAAIIIIAAAggggIAHAgTpHiBRBAEEEEAAAQQQQAABJwUI0p3U5lgIIIAAAggggAACCHggQJDuARJFEEAAAQQQQAABBBBwUoAg3UltjoUAAggggAACCCCAgAcCBOkeIFEEAQQQQAABBBBAAAEnBQjSndTmWAgggAACCCCAAAIIeCBAkO4BEkUQQAABBBBAAAEEEHBSgCDdSW2OhQACCCCAAAIIIICABwIE6R4gUQQBBBBAAAEEEEAAAScFCNKd1OZYCCCAAAIIIIAAAgh4IECQ7gESRRBAAAEEEEAAAQQQcFKAIN1JbY6FAAIIIIAAAggggIAHAgTpHiBRBAEEEEAAAQQQQAABJwUI0p3U5lgIIIAAAggggAACCHggQJDuARJFEEAAAQQQQAABBBBwUoAg3UltjoUAAggggAACCCCAgAcCBOkeIFEEAQQQQAABBBBAAAEnBQjSndTmWAgggAACCCCAAAIIeCBAkO4BEkUQQAABBBBAAAEEEHBSgCDdSW2OhQACCCCAAAIIIICABwIE6R4gUQQBBBBAAAEEEEAAAScFCNKd1OZYCCCAAAIIIIAAAgh4IBDsQRmKIIAAAmUmkH8yTU58NFNOrd8gVZo2latGPSWBNWuWOF5uYoKkTJkiZw8ekLAbb5Kaw0dIUHS0KXfqX99I6pzZUpiTI5Fdukj1Xw6XgMBAObt7l6RMm1asrquefVYqNWos+ceOSco7b0nOzh0SeXMHqXbvvRISd12xsiwggAACCCBwpQRoSb9S8hwXAQRECgvl4K+ekON//rNU799fzh46KHv73y2FBQUldOIH3Sd5GekS/fAjkrNlqyQMftCUy0s+IokPPyxVGjWSanffI8dmz5FjE94w+2dv3iynd/0g4T9r6/4JDI8w2+IHDpDMb9dK9IMPmXp333675J9ILXFcViCAAAIIIHAlBGhJvxLqHBMBBIxA7v5EyVr7rTRZu1aCY+tJxO095Ie2N8jp9d9J6C0d3Eqnt26V/OSjEvvV1xIYWVVCb2gruzt1kjxr/4zly6Xy9dfLVS/9xpQvLCyQI+PGyVW/eUXOWK3vVbvfLtWHDHXXpTP5x1IkNyFRmnz3nQTHXC1VB94nJ/82X7K++EKiBj9UrCwLCCCAAAIIXAkBWtKvhDrHRAABI3A2MVECQkJMgK4rNH1Fl8/uSygmVHgqyywHVK5ifhfm5ZvfZ/bslYKsTAmKCHeXD8jPl4LsbClIz5Az8fGSbgXxuzvdKvuHPCrZa/5lygWGhUvDzz4zAbquyN27Vwpzc6VK6zbuephBAAEEEEDgSgoQpF9JfY6NQAUXOHvwoARG1yimEFz3Gsk/lVlsXZWbb5HAsDBJHjdWcjaul+QJr5vtBZkZEtW3n2Rv3CRpH88xQXjyjBnntlnB+2kr+Nb9Yn79nFSy8twTHnpIsr78UgLCwyW0/c2m3Kl/rpb4u/tKePv2UrlFi2LHZQEBBBBAAIErJUC6y5WS57gIICBBoaFSkHqimESB1fmzUv1ri63TTqCxU96VpJdflBPz5klkj9tNi3tlKw+90nXXSa0nn5TD48ebfXRbxqFDVit5jDRetFQCw8NErNb5yLv7SdamzZK2ZJFE3HGHFFot7kfGPGfSXK56+imp+fSzxY7JAgIIIIAAAldSgCD9SupzbAQquEDI1VebNBMd4SWoejUpyMk2ueeVr4srJlOQli5BVSPl+s3fm86meUePyq727STECtJz9++XqN53yVVjXjDbMq0g/PQPu6Tg9GnJXP6FRN5xpwRGhZj6KjWobx2nuplPenyYnLZGf4n7+mup1Lj48YodnAUEEEAAAQSugADpLlcAnUMigMA5gVBrKEXNQU/94+8l30pPOfr6ayY9Jbh+A9PSffwPM0SHXiwszJd9DzwgWStWWMF3jhx9c4JJT9FOpGd375a9ve+U/JQUyTtyRJLfnSo1BgyQQKuV/uiM6ZL82lgTvGs+etaq1VLtrj6Sl3RAMld+JXWee14CrBFmcvfsNj/6YYEJAQQQQAABXxCgJd0X7gLngEBFFbAC9AYffWRyxY+9/4EJ2K+dO9eMca5jnh+dNEmqWK3cET17mpSWA08+YaQ0j73x54vNfHjXrhJxawfZ1eEW0yqv8zWeeFIkIECuGf+aJD31jJXSUt+UjbKC+SpWLnrW0iVmOelXo8xv1z8xr/xGalhjrDMhgAACCCBwpQUC8vLyCq/0SZTV8dtbHcHWr19fVtVTLwII2CRQmHvWDImo6SsBQUEXrbXQSmHJPXxYKjVsWKJMfmqqCdKD69Qpse3svn0SXCNaAqtFldjGCgQQQAABBHxRgJZ0X7wrnBMCFUwgIKSS6QBa2mUHVKlywQBd93M9ffRCdVwoqL9QOdYhgAACCCDgKwLkpPvKneA8EEAAAQQQQAABBBD4UYAgnZcCAggggAACCCCAAAI+JkCQ7mM3hNNBAAEEEEAAAQQQQIAgndcAAggggAACCCCAAAI+JkCQ7mM3hNNBAAEEEEAAAQQQQIDRXXgNIFBOBVJ/bz2oZ/Lkcnr2nHZZCVS98+cS+8f3y6p66kUAAQQQcEiAlnSHoDkMAggggAACCCCAAAKeChCkeypFOQQQQAABBBBAAAEEHBIgSHcImsMggAACCCCAAAIIIOCpAEG6p1KUQwABBBBAAAEEEEDAIQGCdIegOQwCCCCAAAIIIIAAAp4KEKR7KkU5BBBAAAEEEEAAAQQcEiBIdwiawyCAAAIIIIAAAggg4KmAbeOkb9iwQZYvXy65ubnSrVs36dKlywXPYdOmTbJo0SI5c+aMtG/fXvr37y+BgYGSk5Mjc+fOlR07dkj9+vWlb9++0rhxY1PHN998I2vXri1W30svvSQBAQHF1rGAAAIIIIAAAggggIA/CNjSkr5z504ZNWqUhIaGSocOHWTcuHGyevXqEj7Hjx+XkSNHSr169aR79+4yf/58mTlzpik3Y8YME4j37t1bQkJCZNiwYXLs2DGzbc2aNZKfny/Nmzd3/xCgl+BlBQIIIIAAAggggICfCNjSkr5w4ULp16+fjB492rDs2rXLBOBdu3YtxrRq1Spp3bq1DB8+3KwvLCyUqVOnypAhQ2TZsmUyadIkadeunfTo0UNWrFghGzdulDvvvFMOHDggI0aMkLZt2xarjwUEEEAAAQQQQAABBPxRwJaWdA2imzVr5vaJjY2V+Ph497Jr5tSpUxIeHu5aNK3jJ0+elIyMDMnOzpawsDCzTYN3TZtJSEgwy3v27JE5c+bIwIEDTVC/f/9+dx3MIIAAAggggAACCCDgbwK2BekRERFum9q1a4sG3+dP2rK+bt06WbJkiWzevFnmzZtnimgqS69evWTWrFmydetWk5uu+2dlZUlmZqYJ4K+++mrTmr5v3z6577773Kkw5x+DZQQQQAABBBBAAAEEyruALeku2gKelpbmtjh9+rQ0adLEveyaadCggUl1mTJlilmlQbu2ltesWVOGDh0q48ePN9u1Vb5Fixai5TX419SXqKgos4+mw9x+++2yfv16ueuuu8w6zXXXXPaik9ap05YtW4quZh4BvxEIOXrEb66FC7FPIDM9g/c9+zipCQEEEPhJAtp/UvtW2jHZEqTHxMTIkSP/DRiSkpIkLi6uxPlpmc6dO5uAXFNatHOpBunaCTQ1NVWmTZtmUl50WVNbNG3m8OHDojnuOmKMTpouox8KXEG7rtOAXIP28ycdPaZNmzbnr2YZAb8QSF27RpL94kq4CDsFIqOqSizve3aSUhcCCCBwRQRsSXfRFvEFCxZISkqKbN++3cy3atXKXNC2bdtMS7guJCYmyuDBg+XEiROird865KJ2EtWgfPr06TJ79mzR4H3x4sUmzaVly5ZSpUoVefHFF2XlypUmT12Po/nrN9100xUB46AIIIAAAggggAACCJS1gC0t6Tqyi7Zk9+nTx5yvDq/omtf1OoRiz549zcgtnTp1kgEDBphAW+cHDRpk9tF0Fx3dZenSpWYM9YkTJ0pkZKTZ9tRTT8mECRPkN7/5jVnWoF2DdyYEEEAAAQQQQAABBPxRICAvL6/QrgvTcc31wUTR0dGXrDI9Pd20irvyxl2FCwoKRFNl6tatK0FBQa7V5rd1nnLo0CHR1JpKlSoV23axBU13uVAazMXKsx6B8iSQ+vvpkjx5cnk6Zc7VAYGqd/5cYv/4vgNH4hAIIIAAAmUpYEtLuusEa9Wq5Zq95O+i+eRFC2qAr08bvdAUHBx80W0XKs86BBBAAAEEEEAAAQTKq4AtOenl9eI5bwQQQAABBBBAAAEEfFGAIN0X7wrnhAACCCCAAAIIIFChBQjSK/Tt5+IRQAABBBBAAAEEfFGAIN0X7wrnhAACCCCAAAIIIFChBQjSK/Tt5+IRQAABBBBAAAEEfFGAIN0X7wrnhAACCCCAAAIIIFChBQjSK/Tt5+IRQAABBBBAAAEEfFGAIN0X7wrnhAACCCCAAAIIIFChBQjSK/Tt5+IRQAABBBBAAAEEfFHA1ieO+uIFck4IIIAAAgj8FIFT/1wtJz/7VArOnpVqd98tVfvcXXL3wkI5OecjSf/iSwmqFiXV7xkgEXfcYcoVnjolx6a8I6e+3yyVG8dJ9MOPSuUWLYrVkbtnt+x/8gm59tPPJKhGtEhenhx85uliZard1cfUeXb3LkmZNq3YtquefVYqNWpcbB0LCCDgXwK0pPvX/eRqEEAAAQS8EDj9n82S+OijEhQeLlE9esnBZ/9HMhcvKlFj5pJFkjzpd1K1Zw8J/9nP5MCvRkrW11+bcsljfytpX62UGg88KIEhlSS+392Sd/iwu47C3LOSMHSInNm9RyQ3z6zPPZgkGV9+adXV1v0TUqe22Za9ebOc3vWDe72WCQyPcNfHDAII+KcALen+eV+5KgQQQACByxA4/qdZVut5X4mZ/LbZO3vbFjn2p5kSeXe/YrWd+PRTq4X8IakxdNiP5f5X0hb8TcI73SYnP18k9WfOlPAuXSVqwEBJW7JUsr75p1SzgnadUt6cIJUbNpLc/UlmWf85m5gola+Lk+pDhrrXuWbOJCZI1e63X3Cbqwy/EUDA/wRoSfe/e8oVIYAAAghcpsDpvfFWi/VN7r0rN2pktXjvdS+7ZvJPZRVrzS7Mz5PsnTul0EpbKczNtbaFnytqpcVoy/npH34wy9lr/iUnFi+Wq18d66rK/D5rBeK5Bw/L3jt6Sfydd0j6/M/MfrrxTHy8pC9fLrs73Sr7hzwqWgcTAgj4vwBBuv/fY64QAQQQQMBDgdz9+yUwKtJdutLVdaUgO9u97Jqp0X+gHPt4jmR9+Q+TDpPxjy8kPzNTAkPDJLLH7VYqzETJ+fc6OfbuO2b//PQ0yT95QvYPHy4Nfj+jWICvdZ5JSDRBea0hQ6w0mx5y6IUX5ejr483hTu/dK4FhYRLz6+ekUs2akvDQQ9Zxv3SdCr8RQMBPBUh38dMby2UhgAACCFyGQFio5KWmuncsOJ0tIdc2cC+7ZiIH9JeMNd/I/sdHSFBUNYnoeKuI1WquUx0rwD7wxAjZd//9UqlhQ5PGEnpdE0l+9bcSUr++aW3PtgJ4nbK/WydhHTrKVS+8YH40yNdJA/4Tf/1E6ox7XRovWmoF9db6kBCTdpO1abOVQrPI3VHV7MA/CCDgdwK0pPvdLeWCEEAAAQQuV6By3Wvk7IH97t011SSsWTP3smsmd1+CxIwdJ80T90vTbduk0tVXSxUrENdAPS85WRou/Nxsi1v9T8nLyJDKcXFSWFAo+dZ80pgxcmjCG6aqg+PHy+kd2yXL6mhatHNppbp1JahWtGmFz1z+hfU7x3VoqdSgvgRVr+5eZgYBBPxTgCDdP+8rV4UAAgggcBkCOuzhiT//VfIOHZKcjRskdc5ciWh/i6kpZ/13JldcF9IXLZTEhx8263VEGO0sWvWuu0QCAuTQK69IyltvmYA9bc4cKczIlNCb2ktdK82lybp/m584a+hGna77hzWii9XB9NR36yXRyjcvtIZ9PLt3j5VK87FU73ePlT4TKkdnTJfk18aa+jQfPWvVatHzZEIAAf8WIN3Fv+8vV4cAAggg8BMEajz2C8lcvVp2dTgXmEd07SJRPwbjGau+Nh04o+69T6o/9IhkWK3k/9fsetPafdVToyT0xnMdTmPGPC8Hn3/BGmtdO3/mSr33P5BAayz1S021nrSGcFy3TnbGNTbFNE2m+oMPmaD/mvGvSdJTz8jJv9U326J63ylV2t98qerYhgACfiAQkJeXdy6Jzg8u5vxLaN++vaxfv/781Swj4BcCqb+3WtcmT/aLa+Ei7BOoeufPJfaP79tXYQWtSVNPAoKCJKj2ubHKL8agHU2DatUyHTuLlSkokNz4vRJsDbWo9Xg65R2xxlMPCJTgOnVK7HJ23z4Jth58VFrAX2JHViCAQLkUoCW9XN42ThoBBBBAoCwFgq0cc08m7Qh6wSkwUELirrvgpkutDI65+HG1dZ0JAQQqjgA56RXnXnOlCCCAAAIIIIAAAuVEgCC9nNwoThMBBBBAAAEEEECg4ggQpFece82VIoAAAggggAACCJQTAYL0cnKjOE0EEEAAAQQQQACBiiNAx9GKc6+5UgQQQMAxgaIP5nHsoByoXAh42im3XFwMJ4lAGQoQpJchLlUjgAACFVXgyGvjJOMfX1TUy+e6LyJQ57nnJPpXoy6yldUIIFBUgHSXohrMI4AAAggggAACCCDgAwIE6T5wEzgFBBBAAAEEEEAAAQSKChCkF9VgHgEEEEAAAQQQQAABHxAgSPeBm8ApIIAAAggggAACCCBQVIAgvagG8wgggAACCCCAAAII+IAAQboP3AROAQEEEEAAAQQQQACBogIE6UU1mEcAAQQQQAABBBBAwAcECNJ94CZwCggggAACCCCAAAIIFBUgSC+qwTwCCCCAAAIIIIAAAj4gQJDuAzeBU0AAAQQQQAABBBBAoKgAQXpRDeYRQAABBBBAAAEEEPABAYJ0H7gJnAICCCCAAAIIIIAAAkUFCNKLajCPAAIIIIAAAggggIAPCBCk+8BN4BQQQAABBBBAAAEEECgqQJBeVIN5BBBAAAEEEEAAAQR8QIAg3QduAqeAAAIIIIAAAggggEBRAYL0ohrMI4AAAggggAACCCDgAwIE6T5wEzgFBBBAAAEEEEAAAQSKChCkF9VgHgEEEEAAAQQQQAABHxAgSPeBm8ApIIAAAggggAACCCBQVIAgvagG8wgggAACCCCAAAII+IAAQboP3AROAQEEEEAAAQQQQACBogIE6UU1mEcAAQQQQAABBBBAwAcECNJ94CZwCggggAACCCCAAAIIFBUgSC+qwTwCCCCAAAIIIIAAAj4gQJDuAzeBU0AAAQQQQAABBBBAoKgAQXpRDeYRQAABBBBAAAEEEPABAYJ0H7gJnAICCCCAAAIIIIAAAkUFCNKLajCPAAIIIIAAAggggIAPCBCk+8BN4BQQQAABBBBAAAEEECgqQJBeVIN5BBBAAAEEEEAAAQR8QCDYrnPYsGGDLF++XHJzc6Vbt27SpUuXC1a9adMmWbRokZw5c0bat28v/fv3l8DAQMnJyZG5c+fKjh07pH79+tK3b19p3LhxsTp0nxdffFEefPBBufHGG4ttYwEBBBBAAAEEEEAAAX8RsKUlfefOnTJq1CgJDQ2VDh06yLhx42T16tUljI4fPy4jR46UevXqSffu3WX+/Pkyc+ZMU27GjBmydu1a6d27t4SEhMiwYcPk2LFjxer4wx/+IN9++61kZmYWW88CAggggAACCCCAAAL+JGBLS/rChQulX79+Mnr0aGOza9cuE4B37dq1mNWqVaukdevWMnz4cLO+sLBQpk6dKkOGDJFly5bJpEmTpF27dtKjRw9ZsWKFbNy4Ue68805Tdt26dfL999/LNddcU6xOFhBAAAEEEEAAAQQQ8DcBW1rSDxw4IM2aNXPbxMbGSnx8vHvZNXPq1CkJDw93LUp+fr6cPHlSMjIyJDs7W8LCwsw2Dd41bSYhIcEsa5nx48fLq6++KkFBQe79mUEAAQQQQAABBBBAwB8FbAvSIyIi3D61a9c2wbd7xY8z2rKuLeJLliyRzZs3y7x588wWDdZ79eols2bNkq1bt5rcdA3Ms7KyRAP2N954Qx544AFp1KjR+VWyjAACCCCAAAIIIICA3wnYku6iLeBpaWlunNOnT0uTJk3cy66ZBg0amFSXKVOmmFUatGtrec2aNWXo0KGmtVxTYbRVvkWLFqLltTOqBvQDBw40v7XuvXv3mk6l2mKvk+a6ay570Unr1GnLli1FVzOPgN8IhBw94jfXwoXYJ5CZnuET73uVrPNgQuB8gWTrfSuJv8vns7DsRwLNmzc3fSvtuCRbgvSYmBg5cuS/AUNSUpLExcWVOD8t07lzZxOQawu5di7VID0gIEBSU1Nl2rRpJuVFlzUo1yA8JSVFtGXeFdjr8oIFC6Ry5cryyCOPmGNoQL5+/foSx9PRY9q0aVNiPSsQ8AeB1LVrJNkfLoRrsFUgMqqqxPrA+16SdR6E6bbeWr+orE7tGIn2gdenX2ByEX4vYEu6i7aIa+CsAfT27dvNfKtWrQzetm3bTCdQXUhMTJTBgwfLiRMnTOu3DrmonUQ1KJ8+fbrMnj3bpLcsXrzYjODSsmVL0yH1k08+EdePDss4ZswYd4Du93eIC0QAAQQQQAABBBCocAK2tKTryC7akt2nTx8DqMMruuZ1/Zo1a6Rnz55m5JZOnTrJgAEDTEdRnR80aJDZR9NddHSXpUuXmjHUJ06cKJGRkRXuhnDBCCCAAAIIIIAAAgjYEqQHBwfL5MmTzbjm+mCi6Ohot6yOd64/OunILFouPT3djN7iyhvXbR07djRjrGuqTN26dS86isuf//xnLc6EAAIIIIAAAggggIDfCtgSpLt0atWq5Zq95O+oqKgLbtcAX582yoQAAggggAACCCCAQEUWsCUnvSIDcu0IIIAAAggggAACCNgtQJButyj1IYAAAggggAACCCDgpQBBupeA7I4AAggggAACCCCAgN0CBOl2i1IfAggggAACCCCAAAJeChCkewnI7ggggAACCCCAAAII2C1AkG63KPUhgAACCCCAAAIIIOClAEG6l4DsjgACCCCAAAIIIICA3QIE6XaLUh8CCCCAAAIIIIAAAl4KEKR7CcjuCCCAAAIIIIAAAgjYLUCQbrco9SGAAAIIIIAAAggg4KUAQbqXgOyOAAIIIIAAAggggIDdAgTpdotSHwIIIIAAAggggAACXgoQpHsJyO4IIIAAAggggAACCNgtQJButyj1IYAAAggggAACCCDgpQBBupeA7I4AAggggAACCCCAgN0CBOl2i1IfAggggAACCCCAAAJeChCkewnI7ggggAACCCCAAAII2C1AkG63KPUhgAACCCCAAAIIIOClAEG6l4DsjgACCCCAAAIIIICA3QIE6XaLUh8CCCCAAAIIIIAAAl4KEKR7CcjuCCCAAAIIIIAAAgjYLUCQbrco9SGAAAIIIIAAAggg4KUAQbqXgOyOAAIIIIAAAggggIDdAgTpdotSHwIIIIAAAggggAACXgoQpHsJyO4IIIAAAggggAACCNgtQJButyj1IYAAAggggAACCCDgpQBBupeA7I4AAggggAACCCCAgN0CBOl2i1IfAggggAACCCCAAAJeChCkewnI7ggggAACCCCAAAII2C1AkG63KPUhgAACCCCAAAIIIOClAEG6l4DsjgACCCCAAAIIIICA3QIE6XaLUh8CCCCAAAIIIIAAAl4KBHu5P7uXB4HCQkmb+7GkrVwhIdHREv2LX0qVVq1KnHnhqVNybMo7cur7zVK5cZxEP/yoVG7RwpTLTUyQlClT5OzBAxJ2401Sc/gICbLqEqvuE7M+lIyVX0lwrVpStVdPqdrn7nN15+XJ8d9Pl8y1a6Ry/XpSfcBACb2lg0hAQIljswIBBBBAAAEEEEDgvwK0pP/Xwm/nUt9/Tw6PHy+RN98ilWJjZV//eyQ3YV+J600e+1tJ+2ql1HjgQQkMqSTx/e6WvMOHTbn4QfdJXka6Fbg/IjlbtkrC4AelsKDABOhH335HIjp0kPCbb5akX42S7DX/MvskvzZejr3/vkR16y5VmjSVfQ88IFmrVpU4LisQQAABBBBAAAEEigvQkl7cwy+XUqZOlWvemCBRgwaZ6zu5eJGkLVwgtX79nPt6C3PPysnPF0n9mTMlvEtXibJavdOWLJWsb/4pVZo1l/zkoxL71dcSGFlVQm9oK7s7dZK8/YmmBb3WqFFS88mR5+qeP18yli+XsNs6SfrKlXLNhDek6j0DzLaMNWvk5PzPJKJ7d/dxmUEAAQQQQAABBBAoKUCQXtLEr9YU5GRLQXa2FVjf4L6uyo0ay5n4ePeyzhRaqSmFubkSGB5+br2VxqKB++kffrBSVeqbdQGVq5jfhXn55veZPXul/l8/sbJXAuRs/F7J+e47yfn+e6n9zNNme72p70rlptef20fr2vq/ctVTvzLL/IMAAggggAACCCBwcQGC9Ivb+MUWV7pKYNVI9/VUrhcrZ/YWD9IDQ8MkssftkjxpotT59WjJ+vc6E9znp6dJFStNJjAsTJLHjZVqVgrMsffeM3UVZGZIQOC5jKmjv5skGV8ul4CQEAmueZXZHnpTe/P77L59cuDxYSboj+p/rlXdfTLMIIAAAggggAACCJQQICe9BIl/rQisEmouKD/1pPvCCrNzTMdQ94ofZ+q88KLkp6fLvvvvl7Slf5fK18VJ6HVNTCAeO+VdSfvyC9l370DT8VOD8cqNGrmriH3/Q2m6abNUsTqaHhzzvHt9+p/nyZ6uXaRSgwZy3bp1ElTD6mzKhAACCCCAAAIIIHBJAYL0S/KU/41mBBbrMnKTDrgvJmf3LqnStKl72cxY6S15ycnScOHn0jxxv8St/qfVUTRDKsfFSUFaugRZLfHXb/7ebLv69QkmNSbECtL39ugu2d+uNVUEWaO7VO/bxzrWQbN8cvaf5NCrr0oDK8+93oezJKh6jeLHZAkBBBBAAAEEEEDgggIE6Rdk8Z+VAVWqSHin2yTlvT+a9JX0Tz+1RmfZJmFt20phfr4c/8MM0eEVdVjEQ6+8IilvvWWGVUybM0cKMzJFU1YKC/PPjcyyYoUUnM6Ro29OkPD27U0n0ipNmkjKtGmSu3+/5B06JKmf/U1q3Ge1tlvT0XfelRqDB5tW9Nw9u0V/8o6cGy3Gf4S5EgQQQAABBBBAwH4BctLtN/W5Guv+brLsuau3/N/151rPa48ZI5WsIRELc6yAe9IkqWKNiR7S4FqJsdJUDj7/gpz87DPTUl7v/Q8ksFqUuZ5aTz4pB558wswHRteQxp8vPrf+l8NMAL/b+iCgU4SV2lL9scck/+hRK3UmTVJnzzY/ZuOP2+vP/ti1yG8EEEAAAQQQQACBCwgQpF8Axd9WBcdcLddv+o+c3RcvOq+dQHUKCA2V5vv/mwYTcced0rTnHZJrjdQS3LCRBAQFuSmuGvOC1Hr6Gcm1xk2v1LChe33lNjdI023bJC8xUYLq1DGt666NRet2reM3AggggAACCCCAQOkCBOmlG/lHCSudpZI19GKpkzVaS0jcdRcspqkzRQN0V6EA68FHF9vHVYbfCCCAAAIIIIAAAp4LkJPuuRUlEUAAAQQQQAABBBBwRIAg3RFmDoIAAggggAACCCCAgOcCBOmeW1ESAQQQQAABBBBAAAFHBAjSHWHmIAgggAACCCCAAAIIeC5AkO65FSURQAABBBBAAAEEEHBEgNFdLsGc9fXXsv8XQy5Rgk0VUSA4po40+W5DRbx0rhkBBBBAAAEEHBKgJd0haA6DAAIIIIAAAggggICnAgTpnkpRDgEEEEAAAQQQQAABhwQI0h2C5jAIIIAAAggggAACCHgqQJDuqRTlEEAAAQQQQAABBBBwSIAg3SFoDoMAAggggAACCCCAgKcCBOmeSlEOAQQQQAABBBBAAAGHBAjSHYLmMAgggAACCCCAAAIIeCpg2zjpGzZskOXLl0tubq5069ZNunTpcsFz2LRpkyxatEjOnDkj7du3l/79+0tgYKDk5OTI3LlzZceOHVK/fn3p27evNG7c2NSxf/9++eSTT+Tw4cPStm1b6dmzp8TExFywflYigAACCCCAAAIIIFDeBWxpSd+5c6eMGjVKQkNDpUOHDjJu3DhZvXp1CZvjx4/LyJEjpV69etK9e3eZP3++zJw505SbMWOGrF27Vnr37i0hISEybNgwOXbsmOTl5cljjz0muq8G7nqssWPHlqibFQgggAACCCCAAAII+IuALS3pCxculH79+sno0aONy65du0wA3rVr12JOq1atktatW8vw4cPN+sLCQpk6daoMGTJEli1bJpMmTZJ27dpJjx49ZMWKFbJx40a59tprJTs7WyZOnChBQUFSs2ZNs39GRoZUrVq1WP0sIIAAAggggAACCCDgDwK2tKQfOHBAmjVr5vaIjY2V+Ph497Jr5tSpUxIeHu5alPz8fDl58qRowK2BeFhYmNmmwbumzSQkJEjTpk1l3bp1cvbsWdmyZYtpeW/RogUBuluRGQQQQAABBBBAAAF/E7AtSI+IiHDb1K5d2wTf7hU/zmjLugbcS5Yskc2bN8u8efPMFg3We/XqJbNmzZKtW7ea3HQN3rOysiQgIMC0oOsHgccff1w09/366683Af759bOMAAIIIIAAAggggIA/CNiS7qIt4GlpaW6P06dPS5MmTdzLrpkGDRqYVJUpU6aYVRq0a2u5prAMHTpUxo8fb7Zrq7y2lmt516T1aYCvP5pWozntN9xwg9ms+eqay1500jp10tb3y52CExMk4HJ3Zj+/FcizvuXx5nVlF0zI0SN2VUU9fiSQmZ7hE6/PStZ5MCFwvkCy9b6V5MXf5fPrYxkBXxNo3ry56Vtpx3nZEqTrSCtHjvw3YEhKSpK4uLgS56dlOnfubAJyTWnRzqUapGtreWpqqkybNs2kvOjywIEDRdNmVq5cacpNmDDBtKh37NjRlDl06JA7SNeAfP369SWOp6PHtGnTpsR6T1dkWee039PClKswAsFWx+bmXryu7IJKXbtGku2qjHr8RiAyqqrE+sDrM8k6D8J0v3lZ2XYhdWrHSLQPvD5tuyAqQqAMBWxJd9EW8QULFkhKSops377dzLdq1cqc9rZt20wnUF1ITEyUwYMHy4kTJ8xoLTrkonYS1aB8+vTpMnv2bNHgffHixZKZmSktW7Y0ue5ff/216I/mpX/55Zcmf12DdSYEEEAAAQQQQAABBPxRwJaWdB3ZRVuy+/TpY4w0FcU1r+vXrFljxjbXkVs6deokAwYMMIG2zg8aNMjso+kuOrrL0qVLzRjqOppLZGSk+dH6X3rpJVPummuukTFjxki1atX88X5wTQgggAACCCCAAAIIiC1BenBwsEyePNmMa64PJoqOjnbT6njn+qOTDqGo5dLT083oLa68cd2mLeM6xrqmytStW9eU1fU6vfjiizJixAjR0WF0GxMCCCCAAAIIIIAAAv4sYEuQ7gKqVauWa/aSv6Oioi64XQN8fdrohabq1auL/jAhgAACCCCAAAIIIODvArbkpPs7EteHAAIIIIAAAggggICTAgTpTmpzLAQQQAABBBBAAAEEPBAgSPcAiSIIIIAAAggggAACCDgpQJDupDbHQgABBBBAAAEEEEDAAwGCdA+QKIIAAggggAACCCCAgJMCBOlOanMsBBBAAAEEEEAAAQQ8ECBI9wCJIggggAACCCCAAAIIOClAkO6kNsdCAAEEEEAAAQQQQMADAYJ0D5AoggACCCCAAAIIIICAkwIE6U5qcywEEEAAAQQQQAABBDwQIEj3AIkiCCCAAAIIIIAAAgg4KUCQ7qQ2x0IAAQQQQAABBBBAwAMBgnQPkCiCAAIIIIAAAggggICTAgTpTmpzLAQQQAABBBBAAAEEPBAgSPcAiSIIIIAAAggggAACCDgpQJDupDbHQgABBBBAAAEEEEDAAwGCdA+QKIIAAggggAACCCCAgJMCBOlOanMsBBBAAAEEEEAAAQQ8ECBI9wCJIggggAACCCCAAAIIOClAkO6kNsdCAAEEEEAAAQQQQMADAYJ0D5AoggACCCCAAAIIIICAkwIE6U5qcywEEEAAAQQQQAABBDwQIEj3AIkiCCCAAAIIIIAAAgg4KUCQ7qQ2x0IAAQQQQAABBBBAwAMBgnQPkCiCAAIIIIAAAggggICTAgTpTmpzLAQQQAABBBBAAAEEPBAgSPcAiSIIIIAAAggggAACCDgpQJDupDbHQgABBBBAAAEEEEDAAwGCdA+QKIIAAggggAACCCCAgJMCBOlOanMsBBBAAAEEEEAAAQQ8ECBI9wCJIggggAACCCCAAAIIOClAkO6kNsdCAAEEEEAAAQQQQMADAYJ0D5AoggACCCCAAAIIIICAkwIE6U5qcywEEEAAAQQQQAABBDwQIEj3AIkiCCCAAAIIIIAAAgg4KUCQ7qQ2x0IAAQQQQAABBBBAwAMBgnQPkCiCAAIIIIAAAggggICTAgTpTmpzLAQQQAABBBBAAAEEPBAgSPcAiSIIIIAAAggggAACCDgpQJDupDbHQgABBBBAAAEEEEDAAwGCdA+QKIIAAggggAACCCCAgJMCBOlOanMsBBBAAAEEEEAAAQQ8ECBI9wCJIggggAACCCCAAAIIOClAkO6kNsdCAAEEEEAAAQQQQMADAYJ0D5AoggACCCCAAAIIIICAkwIE6U5qcywEEEAAAQQQQAABBDwQIEj3AIkiCCCAAAIIIIAAAgg4KUCQ7qQ2x0IAAQQQQAABBBBAwAMBgnQPkCiCAAIIIIAAAggggICTAsFOHoxjIYAAAggggAACCFymQGGhpM39WNJWrpCQ6GiJ/sUvpUqrViUqKzx1So5NeUdOfb9ZKjeOk+iHH5XKLVqYcrmJCZIyZYqcPXhAwm68SWoOHyFBVl2uKf9Yihx4YoTUnztPAkPDzOpL1Zf59yWS/uVy1+7md93pvxcJCCi2joWfLkBL+k83Yw8EEEAAAQQQQMBxgdT335PD48dL5M23SKXYWNnX/x7JTdhX4jySx/5W0r5aKTUeeFACQypJfL+7Je/wYVMuftB9kpeRbgXuj0jOlq2SMPhBKSwokIKcbEn75K+yf9gwyd64SeRsnrveS9WX/uWXUpifJ+E/a+v+IUB303k1Q0u6V3zsjAACCCCAAAIIOCOQMnWqXPPGBIkaNMgc8OTiRZK2cIHU+vVz7hMozD0rJz9fJPVnzpTwLl0lasBASVuyVLK++adUadZc8pOPSuxXX0tgZFUJvaGt7O7USfL2J0pAaKikL//SXY9r5lL1VbM+BJzeGy9Xv/yyhN3WybULv20SoCXdJkiqQQABBBBAAAEEykpAW7oLsrOtwPoG9yEqN2osZ+Lj3cs6U5iXJ4W5uRIYHn5uvZUio4H26R9+kMJTWWZdQOUq5ndhXr75fWbPXgmuEyP1P5pjfQh489x+P/57qfq0yNm9e+Xou+/ID+1vkiMvvyi5e3YX25+FyxcgSL98O/ZEAAEEEEAAAQQcEXClqwRWjXQfr3K9WMlPT3cv64zmkUf2uF2SJ02UnH+vk2NWAK3BfX56mlSx0mQCw8IkedxYydm4XpInvG72LcjMKFZH0YVL1VeQlm4+EFSKrSdXv/CC9UHg/2T37be7U2uK1sP8TxcgSP/pZuyBAAIIIIAAAgg4KhBYJdQcLz/1pPu4hdk5pmOoe8WPM3VeeNEE7/vuv1/Slv5dKl8XJ6HXNZGAwECJnfKupH35hey7d6Dp3BkQEiKVGzU6v4piyxerLzCqqjT9z/dyzbvTpOo9A6TezD+Z/U6tXl1sfxYuT4Cc9MtzYy8EEEAAAQQQQMAxAdcILLlJB6Ry82bmuDm7d0mN+x8sfg5WektecrI0XPi5BEZYre7WKCuailI5Lk605TvIaom/fvP3Vl6MVe7oUdnVvp2EXCpIv0R9uVYu++mtWyTy7nvMOQRGVRMN+gOrVy9+TixdlgAt6ZfFxk4IIIAAAggggIBzAgFVqkh4p9sk5b0/mvSV9E8/tUZn2SZhbduak8iwOpDmrPvWBOWHXnlFUt56ywTiaXPmSGFGpoTe1N6Ky/Nl3wMPSNaKFVJwOkeOvjlBwtu3N51IL3olVpB/sfo0FebAU0+LHlvz3k/M/MCkv4R37XLR6tjguYBtLekbNmyQ5cuXS67VWaFbt27SpcuFb9CmTZtk0aJFcubMGWlvvTD69+8vgdbXLzk5OTJ37lzZsWOH1K9fX/r27SuNGzc2V3Ly5EmZY73I9uzZIzdYHSZ69Ohhynh+mZREAAEEEEAAAQTKt0Dd302WPXf1lv+7vqm5kNpjxkilJufmUz74QCKsuCq0w60SM+Z5Ofj8C3Lys89M0Fzv/Q8ksFqU2afWk0/KgSefMPOB0TWk8eeLL4xSpBn34vVFSZ1f/9o61vNS+Oyzpp6rx77qHl/9whWz1lMBW4L0nTt3yqhRo2TgwIFy0003ybhx46xPa4XStWvXYudx/PhxGTlypAwdOlTq1asns2fPlhMnTsjw4cNlxowZsm3bNnnooYdk165dMswap/Mz68VVq1YtefrppyU/P9+97b777pOVK1dK1apVi9XPAgIIIIAAAggg4K8CwTFXy/Wb/iNn98WLzmsnUNfUuMgDhSLuuFOa9rxDcuOtUVsaNpKAoCBXMblqzAtS6+lnJNcaN71Sw4bu9a4ZTaVpvv+Aa9H8vlR90VZLevQTT8pZK/UlpG6saIs/kz0CtgTpCxculH79+sno0aPNWWmQPX/+/BJB+qpVq6R169YmKNeCGshPtcb8HDJkiCxbtkwmTZok7dq1My3lK6yvYjZu3Ci33HKLCdp1e82aNaVXr16yePFi+de//iV33XWXPQrUggACCCCAAAIIlAcBK/2kkjX0YqmTlaUQEnfdBYtpIH2hAP2ChV0rL1GfWHnolRrHuUry2yaBIl9mXH6NBw4ckGbNznVi0FpiradgxZ83bqeuP2U9pjbcNW6ntayt45rKkpGRIdnW8EBhP34i1OBd02YSEhKkivVC+vDDD02ArnXosbRs06bnvt7RdUwIIIAAAggggAACCPiTgG1BekREhNuldu3aJvh2r/hxRtNf1q1bJ0uWLJHNmzfLvHnzzBYN1rWFfNasWbJ161aTm67Be1ZWloRaT8Bq1aqVKad57yNGjJBbb73Vna9+/jFYRgABBBBAAAEEEECgvAvYku6iLeBpaWlui9OnT0uTJk3cy66ZBg0amFSXKVOmmFUatGtruaaxaJ76+PHjzXZtlW/RooVoeZ00iH/nnXdMCs2TVocHzVsvOmmue+/evYuucre8b9mypdj6n7IQnJggAT9lB8pWCIE861seb15XdiGFHD1iV1XU40cCmekZPvH6rGSdBxMC5wskW+9bSV78XT6/PpYR8DWB5s2bW9k/Ibacli1BekxMjBw58t+AISkpSeKs8TjPn7RM586dTUCuKS2rrcHuNUgPsPKrUlNTZdq0aSblRZe1E6qmzeikHVE1fUY7kurIL+dPGuSvX7/+/NVm9Jg2bdqUWO/piizrnPZ7WphyFUYg2PrP19yL15VdUKlr10iyXZVRj98IRFoPF4n1gddnknUehOl+87Ky7ULq1I6RaB94fdp2QVSEQBkK2JLuoi3iCxYskJSUFNm+fbuZd6Wo6Igt2glUp8TERBk8eLAZ0UVbv3XIRR1OUYPy6dOny2xrtBcN3rVjaGZmprRs2VKSrQH5dWjHX/7yl6aO/fv3i/7odiYEEEAAAQQQQAABBPxRwJaWdB3ZRVuy+/TpY4y6d+/untf1a9askZ49e5qRWzp16iQDBgwwnT91ftCgQWYfTXfR0V2WLl1qxlCfOHGiREZGio6rrtMLL7xgfrv+ee655+Tee+91LfIbAQQQQAABBBDwSCD5ty97VI5CFUsgonM3ibBiWF+ZbAnSg4ODZfLkyXLs2DHzYKLo6Gj39el45/qjU5A1TqeWS09PN6O3aJqKa+rYsaN06NBBNFWmbt26pqxu01b6C6WyuPbjNwIIIIAAAggg8FME0leslLwjJAz+FLOKUFaDdF+abAnSXRekDx7yZIqKOvfUq/PL6pNHL5Rzfn45lhFAAAEEEEAAAQQQ8GcBW3LS/RmIa0MAAQQQQAABBBBAwGkBgnSnxTkeAggggAACCCCAAAKlCBCklwLEZgQQQAABBBBAAAEEnBYgSHdanOMhgAACCCCAAAIIIFCKAEF6KUBsRgABBBBAAAEEEEDAaQGCdKfFOR4CCCCAAAIIIIAAAqUIEKSXAsRmBBBAAAEEEEAAAQScFiBId1qc4yGAAAIIIIAAAgggUIoAQXopQGxGAAEEEEAAAQQQQMBpAYJ0p8U5HgIIIIAAAggggAACpQgQpJcCxGYEEEAAAQQQQAABBJwWIEh3WpzjIYAAAggggAACCCBQigBBeilAbEYAAQQQQAABBBBAwGkBgnSnxTkeAggggAACCCCAAAKlCBCklwLEZgQQQAABBBBAAAEEnBYgSHdanOMhgAACCCCAAAIIIFCKAEF6KUBsRgABBBBAAAEEEEDAaQGCdKfFOR4CCCCAAAIIIIAAAqUIEKSXAsRmBBBAAAEEEEAAAQScFiBId1qc4yGAAAIIIIAAAgggUIoAQXopQGxGAAEEEEAAAQQQQMBpAYJ0p8U5HgIIIIAAAggggAACpQgQpJcCxGYEEEAAAQQQQAABBJwWIEh3WpzjIYAAAggggAACCCBQigBBeilAbEYAAQQQQAABBBBAwGkBgnSnxTkeAggggAACCCCAAAKlCBCklwLEZgQQQAABBBBAAAEEnBYgSHdanOMhgAACCCCAAAIIIFCKAEF6KUBsRgABBBBAAAEEEEDAaQGCdKfFOR4CCCCAAAIIIIAAAqUIEKSXAsRmBBBAAAEEEEAAAQScFiBId1qc4yGAAAIIIIAAAgggUIoAQXopQGxGAAEEEEAAAQQQQMBpAYJ0p8U5HgIIIIAAAggggAACpQgQpJcCxGYEEEAAAQQQQAABBJwWIEh3WpzjIYAAAggggAACCCBQigBBeilAbEYAAQQQQAABBBBAwGkBgnSnxTkeAggggAACCCCAAAKlCBCklwLEZgQQQAABBBBAAAEEnBYgSHdanOMhgAACCCCAAAIIIFCKAEF6KUBsRgABBBBAAAEEEEDAaQGCdKfFOR4CCCCAAAIIIIAAAqUIEKSXAsRmBBBAAAEEEEAAAQScFiBId1qc4yGAAAIIIIAAAgggUIoAQXopQGxGAAEEEEAAAQQQQMBpAYJ0p8U5HgIIIIAAAggggAACpQgQpJcCxGYEEEAAAQQQQAABBJwWIEh3WpzjIYAAAggggAACCCBQigBBeilAbEYAAQQQQAABBBBAwGkBgnSnxTkeAggggAACCCCAAAKlCBCklwLEZgQQQAABBBBAAAEEnBYgSHdanOMhgAACCCCAAAIIIFCKAEF6KUBsRgABBBBAAAEEEEDAaQGCdKfFOR4CCCCAAAIIIIAAAqUIEKSXAsRmBBBAAAEEEEAAAQScFiBId1qkWWySAAAiCklEQVSc4yGAAAIIIIAAAgggUIpAcCnbPd68YcMGWb58ueTm5kq3bt2kS5cuF9x306ZNsmjRIjlz5oy0b99e+vfvL4GBgZKTkyNz586VHTt2SP369aVv377SuHHjYnV89tlnUqVKFbOt2AYWEEAAAQQQQAABBBDwIwFbWtJ37twpo0aNktDQUOnQoYOMGzdOVq9eXYLp+PHjMnLkSKlXr550795d5s+fLzNnzjTlZsyYIWvXrpXevXtLSEiIDBs2TI4dO2a27dq1S+bMmSNvv/22JCUllaiXFQgggAACCCCAAAII+JOALS3pCxculH79+sno0aONjQbVGoB37dq1mNWqVaukdevWMnz4cLO+sLBQpk6dKkOGDJFly5bJpEmTpF27dtKjRw9ZsWKFbNy4Ue68807Ztm2b7N69W6pXr16sPhYQQAABBBBAAAEEEPBHAVta0g8cOCDNmjVz+8TGxkp8fLx72TVz6tQpCQ8Pdy1Kfn6+nDx5UjIyMiQ7O1vCwsLMNg3eNW0mISHBLA8cOFDeeOMNadu2rXtfZhBAAAEEEEAAAQQQ8FcB24L0iIgIt1Ht2rVN8O1e8eOMtqyvW7dOlixZIps3b5Z58+aZLRqs9+rVS2bNmiVbt241uekavGdlZZ1fBcsIIIAAAggggAACCPi9gC3pLtoCnpaW5sY6ffq0NGnSxL3smmnQoIFJdZkyZYpZpUG7tpbXrFlThg4dKuPHjzfbtVW+RYsWouU9mTTXXXPZi05ap05btmwpuvonzQcnJkjAT9qDwhVBIM/6lseb15VdRiFHj9hVFfX4kUBmeoZPvD4rWefBhMD5AsnW+1aSF3+Xz6/vcpdDrPdxJgTOF0i04r48L1+fzZs3N30rz6/7cpZtCdJjYmLkyJH/BgzauTMuLq7E+WiZzp07m4BcU1q0c6kG6QEBAZKamirTpk0zKS+6rCkumjbjyaQB+fr160sU1dFj2rRpU2K9pyuyrHPa72lhylUYgWCrY3NzL15XdkGlrl0jyXZVRj1+IxAZVVVifeD1mWSdB2G637ysbLuQOrVjJNoHXp+7rPfxPNuuior8RaBBg2slwgdeny5PW9JdtEV8wYIFkpKSItu3bzfzrVq1MsfQTp/aCVSnxMREGTx4sJw4cUK09VuHXNROohqUT58+XWbPni0avC9evFgyMzOlZcuWZj/+QQABBBBAAAEEEECgIgnY0pKuI7toS3afPn2MnQ6v6JrX9WvWrJGePXuakVs6deokAwYMMB1FdX7QoEFmH0130dFdli5dasZQnzhxokRGRha7FxrM6w8TAggggAACCCCAAAL+LGBLkB4cHCyTJ08245rrg4mio6PdZjreuf7oFBQUZMqlp6eb0VtceeO6rWPHjmaMdU2VqVu3rimr64tOOsILEwIIIIAAAggggAAC/i5gS5DuQqpVq5Zr9pK/o6KiLrhdA3x92igTAggggAACCCCAAAIVWcCWnPSKDMi1I4AAAggggAACCCBgtwBBut2i1IcAAggggAACCCCAgJcCBOleArI7AggggAACCCCAAAJ2CxCk2y1KfQgggAACCCCAAAIIeClAkO4lILsjgAACCCCAAAIIIGC3AEG63aLUhwACCCCAAAIIIICAlwIE6V4CsjsCCCCAAAIIIIAAAnYLEKTbLUp9CCCAAAIIIIAAAgh4KUCQ7iUguyOAAAIIIIAAAgggYLcAQbrdotSHAAIIIIAAAggggICXAgTpXgKyOwIIIIAAAggggAACdgsQpNstSn0IIIAAAggggAACCHgpQJDuJSC7I4AAAggggAACCCBgtwBBut2i1IcAAggggAACCCCAgJcCBOleArI7AggggAACCCCAAAJ2CxCk2y1KfQgggAACCCCAAAIIeClAkO4lILsjgAACCCCAAAIIIGC3AEG63aLUhwACCCCAAAIIIICAlwIE6V4CsjsCCCCAAAIIIIAAAnYLEKTbLUp9CCCAAAIIIIAAAgh4KUCQ7iUguyOAAAIIIIAAAgggYLcAQbrdotSHAAIIIIAAAggggICXAgTpXgKyOwIIIIAAAggggAACdgsQpNstSn0IIIAAAggggAACCHgpQJDuJSC7I4AAAggggAACCCBgtwBBut2i1IcAAggggAACCCCAgJcCBOleArI7AggggAACCCCAAAJ2CxCk2y1KfQgggAACCCCAAAIIeClAkO4lILsjgAACCCCAAAIIIGC3AEG63aLUhwACCCCAAAIIIICAlwIE6V4CsjsCCCCAAAIIIIAAAnYLEKTbLUp9CCCAAAIIIIAAAgh4KUCQ7iUguyOAAAIIIIAAAgggYLcAQbrdotSHAAIIIIAAAggggICXAgTpXgKyOwIIIIAAAggggAACdgsQpNstSn0IIIAAAggggAACCHgpQJDuJSC7I4AAAggggAACCCBgtwBBut2i1IcAAggggAACCCCAgJcCBOleArI7AggggAACCCCAAAJ2CxCk2y1KfQgggAACCCCAAAIIeClAkO4lILsjgAACCCCAAAIIIGC3AEG63aLUhwACCCCAAAIIIICAlwIE6V4CsjsCCCCAAAIIIIAAAnYLEKTbLUp9CCCAAAIIIIAAAgh4KUCQ7iUguyOAAAIIIIAAAgggYLcAQbrdotSHAAIIIIAAAggggICXAgTpXgKyOwIIIIAAAggggAACdgsQpNstSn0IIIAAAggggAACCHgpQJDuJSC7I4AAAggggAACCCBgtwBBut2i1IcAAggggAACCCCAgJcCBOleArI7AggggAACCCCAAAJ2CxCk2y1KfQgggAACCCCAAAIIeClAkO4lILsjgAACCCCAAAIIIGC3AEG63aLUhwACCCCAAAIIIICAlwIE6V4CsjsCCCCAAAIIIIAAAnYLEKTbLUp9CCCAAAIIIIAAAgh4KUCQ7iUguyOAAAIIIIAAAgggYLdAsN0Vnl9fYWGhLF26VNatWyfVq1eXfv36SZMmTc4vJjk5OTJ37lzZsWOH1K9fX/r27SuNGzc25Q4dOiQff/yxJCcnS4sWLeS+++6TqKioEnWwAgEEEEAAAQQQQAABfxAo85b0v/3tbzJlyhRp3bq11K5dW0aMGCEHDx4sYTdjxgxZu3at9O7dW0JCQmTYsGFy7NgxU+6ZZ56RrKwsE7jv2rVLnnvuOSkoKChRBysQQAABBBBAAAEEEPAHgTJvSf/Tn/4kY8aMkTvuuMN4rVy5Ur766it57LHH3H65ubmybNkymTRpkrRr10569OghK1askI0bN0qjRo3kwIEDMnv2bAkPD5frr79e7rnnHjl8+LDUrVvXXQczCCCAAAIIIIAAAgj4i0CZtqSfPn1aTp48KU2bNnV7NWjQQJKSktzLOpOfny/Z2dkSFhZm1muKjAbuCQkJJg1GV2rruk6uFnQN3JkQQAABBBBAAAEEEPBHgTIN0l3pKhEREW67mJgYk7riXmHNVKlSRXr16iWzZs2SrVu3mtx0De41xaVVq1Yml/29996T7du3i/7W6dSpU+Y3/yCAAAIIIIAAAggg4G8CAXl5eYVldVEpKSnSp08f+ctf/mLSVvQ4b731loSGhsrIkSOLHXb//v0yfvx4E4g3a9ZMAgMDpWfPnjJo0CD59ttv5e233xbtQKrB/Jo1a0yw7uqAevz4cZPLXrRCDfy1JZ8JAQQQQAABBBBAAAGnBNavX2/Loco0J71atWrmJHVUFs0t10lTWLRzaNFJ01tSU1Nl2rRpJuUlICBABg4cKLGxsZKZmWly0RcuXCiucsuXLy+Wj16zZk25EEj79u0vuL7osZlH4EoI8Nq8Euoc01MBXp+eSlHuSgjw+rwS6hzTUwF9fdo1lWm6S6VKlaRr167yySefmNzyL7/8UjZt2mQ6f+oFaCfSLVu2iAbl06dPN51DNRBfvHixCc5btmxprvPxxx83QzieOXNGPvzwQ7n11ltN4G4XAvUggAACCCCAAAIIIOBLAmXakq4XOnr0aBk+fLh06dLFXPezzz4r1157rZnX4Rl1aMY2bdrI0KFDzeguOqa6BuMTJ06UyMhIU073f/nll828tpr/4Q9/MPP8gwACCCCAAAIIIICAPwqUeZCuQfWCBQvM2Og6r/norumDDz5wzUrHjh2lQ4cOZuQXHVoxKCjIvU0D+Icfflg0x51hF90sZT6jI+lo3wAmBBBAAAEEEPAPAU0n1pQMbURl8m2BMu046tuXztldSmDbtm3mmw0dn14/MBGsX0qLbQgggAACCJQPgXfffdekIX/xxRdm9LzycdYV8yxpJq2Y973Uq9ax7XNyckwa0vvvv19qeQog4LSAdkh/+umnGcXJaXiO55HA3r17zYhlHhWmEAIOCjz11FNmkA5XGrGDh+ZQP1GAIP0nglWU4jrGvY5Vv2vXLvOE14py3Vxn+RGoVauWHD161LQIlZ+z5kz9XUAHQxgyZIgMGzbMPElbhwxmQsBXBF588UWTWqwPkNy8ebN5No2vnBvnUVKAIL2kSYVdoy3nq1atkvnz58tVV10ln3/+ubRr107mzJlTYU24cN8S0JGi/vznP5uT0jSs5557Tv74xz+a/iq+daacTUUU0MYNHRxB+1FpKkH37t1lypQpcvbs2YrIwTX7mMDvf/978zf+d7/7ncybN888yf2VV17xsbPkdIoKEKQX1ajA8/pAqMGDB8uKFSvMH5df/epXUrVqVXnmmWdEx6jfs2dPBdbh0n1FoF69euZ5Cvp61elnP/uZaRVixCdfuUMV+zz0ffLqq682AyHoA/X0AX06FLGOZMaEwJUW2LlzpzRu3Fg6depkfv/pT38y30Z+9tlnV/rUOP5FBAjSLwJT0Vbrg6T0aa469KU+DVa/stWHRulDqO677z7zpFh94itPca1or4wre70a9OhTil944QXzxGEdAUp/tPXcNelTjbXVcvv27a5V/EbAEYENGzaIPsdDhwk+cuSIREdHi+aia6qgTsHBwdKjRw/zwVIf2MeEgFMC1tPk5aWXXpK77rpLRowYIfotjzZq6NPdXdN1111nnjmjz6nh2x6Xim/9Jkj3rfvh2NloS6QGNv/5z3/MMXX0Fk1t+eabb8x/bA3Mp06dajqPam6l/gfv37+/7N6927Fz5EAVW+D77783wU+DBg1M+pWmEGhKlnYW/fvf/y46ApFOJ06cMNv1YWlMCDghoO+f+t752muvmT47Oryw6xkgTZo0MSku+fn55inZ+kHTlT7oxLlxDARU4JFHHpHvvvtObrvtNvPB8d5775XOnTtLbm6uSRPUMhkZGea5NJo6SN8JFfG9iSEYfe+elPkZ6R+XX//61yYo/+c//ymvvvqq3HHHHfK///u/8pvf/MY8/VW/stX/3PrEWA3Or7/+evMHR1NgmBBwQkBbgfSDY79+/WTRokXy5ptvms54v/zlL2XWrFkmp1L/6GiwrukutWvXNk8vduLcOEbFFtDXm/bV0XxebanUVsj7779fNBDSJ2I/9thjoiNk6aRpLw888ID5llJTB5kQKGsBfbK7fiOu/ctq1Kghzz//vHnau34jqYNBaH+zOnXqSHp6uugHTE13YZjlsr4rl1d/mT/M6PJOi73KSkC/AtMAfdKkSSYA0mHs9D+p/gf96quvzFez2nK5cuVK0+mpevXq5j+56+mvZXVe1IuAfmDUPy76wVC/ltUgR59IrF/FaovQ66+/bj5EalD0i1/8QnR0l8zMTNH+E/oaZkKgrASysrJk48aNoikrXbp0MQ/X0zxzV/qf5p1rS7o+HEYbPDQY//bbb01+ur6GNTBiQqCsBDTXXAPtatWqmQ+I2kKukwbjn376qXkIpI6NrmmCH3/8sdx8882iT3fXv/WarkWAXlZ3xvt6aUn33tDna9AcyYMHD5r/mIWFheaPjLZMxsTEmHPXQEdz0LU1SFsvNUjStJb33nvPfE3r8xfICZZrAU1h0T8mS5YskRYtWpg/HvqHRNMG9A+JvlZ1VIKAgADzde0111xjgqR77rmnXF83J18+BDS1RXPONX9X3yM1+NZWdA28NejR16e2lus0atQoE5jrMHdbt26VtWvXmgBp5syZJoDXb36YELBTQBs3tIVcv/3WtFQNuHXkFn2qqKax/Pa3vzX9zfS99I033jANHnYen7rKViDIuoFjy/YQ1H4lBfSPg/7H1FzzH374QXr27Gk6hEZERJiASM9NA3Rt+dHOJTfccIP5NK4PO9BOUEwIlKWAtkzefffdkpCQIPp029tvv920TupQi7pev93RAEg/OOpDtbQFSF/DOjqBtl4yIVDWAjpCi3ZW1uE+tXO9tlLqa1UbNHQ0LG1lb9u2rTmNZs2amWBJA6awsDAzDnVaWppprbzpppvK+lSpvwIK6Khs+rrUD4yuRo59+/ZJmzZtzAdJTXc5deqUiQM0ffDnP/95BVQqv5dMukv5vXeXPHNtxdGvv+Lj4+Uf//iH6XD34IMPmpZJzekdN26c6B+UVq1ayfr1680oBFrhjTfeeMl62YiAtwLacq7f1OhrTz8IamCuD9UIDw83VWsH0WXLlpnXrQbnms6i4/drCox+4KxcubK3p8D+CFxSQFvKNdWvb9++pnWyaAu4tqrre6umGGh6izZuaAqW9onQ9AH90ck1hK1Z4B8EbBLQ90J97T300EOmQU1TWOPi4kQb1nS0IX0/jY2NNY0c+s354sWLzbeU+iFRx0dnKl8CpLuUr/vl8dnqEGDa4VNbI3Wsc500z3zChAnmP61+ZTt27Fjzh0jTXnQkFzqFesxLQQ8E9GmgmirQvHlzd+m//vWv8sEHH5hAW4f31Ae9aMtk7969RXMmb7nlFlNWRx7SPyj6Fa2ORqBl6tev766HGQS8FdA0QM3hvVB/htWrV5sxzjUFS1OtNCVQ3ztdk36Q1I6iOryijjSkQTz9dlw6/C4rgSeffNI9Ipt+iNQAXFvR9emhOv65vkY1QNdvfiZPnmwGfyirc6FeZwQYgtEZ5zI/igZDOuKFtuxohyZNZ3niiSfk66+/do9/qp+w9RO3PsBAv/LS1kpNIfjoo48I0Mv8DlW8A+goQpor6Rp/91//+pf85S9/kQULFpgcdO0kqp2aNEjSXN533nlHtFVIJ/3Do69XTRXQP0YE6BXv9VPWV6zvexqA66Tvn/p+qcG4Tto5VHPQ9T1Vc3v1fVRHytBJv53UdALdrpO2ohOgGwr+KSMBfc/Ub3V0JCv9ZlyfuqwNcToEqL4+derYsaN5n9SRXDQX3ZWCVUanRLUOCRCkOwRdlofRcaL1D4yONNCtWzfTe1tHb9GWdM3ndT3tTjve/c///I9osHTmzBkTHBH8lOWdqdh1a6CtAbjml+ukLev6mPTQ0FAzupCmW2mQpAGSjsuv40prJzydtPPTyy+/bDrdmRX8g4DNAjpEnX5jo4GP5u3q+6fmmuuk75X6wVG/+dEHEml6oLZM6nuqpgtqiyXvnYaKfxwQ0FS/lJQUadiwoXmtuh4yqN+O67flGqDra1f77Pz73/82nUVd6YMOnB6HKEMB0l3KENepqvU/p47eomOc66StPDour7YCacuQprvop29XR1ANhvSTNhMCZSWgrefaWVlzJDVlQL+10cBHPzTqB0UdXlFzKDXg0dFatPOTftjUFIRHH320rE6LehEoJqCpVxrU6Hul5vlqGovmm7sCcG2lrFu3rknL0jGl9amiGiCFhIQUq4cFBMpaQEcM0pFctDO9BuAFBQXm20Z9fepoWIcPHzb9J3TITyb/EaAlvRzeSx2lRfN39XHpOuSSBkT6B8Q16adtzUnTFnPN8dWHa6xbt861mQDdLcFMWQjoh0AdM1pb0HWUC5105AH9kKhBuwbjul1zzfW1e+DAAfNa1dcpAXpZ3BHqvJjAww8/bAJvfVqtjoyhaYD63qoBkHZw1teyPkb90KFDEhUVZT5cEqBfTJP13gpoup8OR1t00m96NEDXhw7qN4w6pKJOOq+pLZp+pX129L2WAL2onH/MM7pLObuPOmqLtozrw1x0XodY1BQCbQnS4Ee/ttVJnxCq/3F10g54+pUtEwJ2CWhHJX1NXWgYxO3bt5tWSR3WU/PJddgvfb1qqoA+Hl2Dcs1N11YhfZqoK6fSrnOjHgQ8FdDUq6efftp0nNfOnzqSkI6aoQ9801bzPn36mMer882jp6KU80ZAn/o9ceJE07Feh5rVbxt15Ctt4NAPjPpNj/Yj028ctaOoPjhLU1wv9D7szXmwr+8IkO7iO/fCozPR4b80P02fEqbDK2kKgbb6aP6kjjmtueia+6sdSvRHx0plQsBuAc3n1W9pNKDR15uOcKGTpq1ocKMjX6xZs8b9x0NHEkpKSpIPP/zQdMLTr2y7WJ3zGLPX7jtDfT9VQN8/NfjRD5Pat0dbzjXXVxs69NsdJgTKUkCHo9UGNdfwx/pa1OdG6HvksGHDzPukfuPjmjQw19flK6+84lrFbz8WoHm1nN1cbdFxDRmmAbqmC7z99tumI5P+1g542iFP/wMToJezm1uOTveRRx4xrTw69q6OKKQdl3S0Fu0PoR8O9UEuOv7+bbfdZq5KRxXSTnr6h0dbiHToOiYEfEFA0wb0Nazf9mjLueaja18JJgScEHj99ddNnzJ9b9TXonZK1s6gOjytfkN+/lT0Cbfnb2PZ/wRoSS9n91Q7NenYqPqwDU030BEJ9Gva7777zrSqu1rXy9llcbrlUEADG/0qVlvTtfUnOTnZ/HHRvhLab0JTsTS/1zXEoqYVtGzZkhFbyuG9rginrB8i9cMmwylWhLt95a5Rv3nUv+P6fql9cDTlSlP+9NtH1zNN9JtKfW/V16SmDDJVXIEgqxPC2Ip7+eXvyrVziD7gRfPR9SExOl6qBuqa26st6/rDhEBZCmiHT/1GR9MB9PWnrY/aUq5j8+sfHB1/Woes0w+RGqTro9P1qYxajgdmleWdoW5vBPQ1zNNsvRFk39IENLVFn/ytHZH1R/vm6N9xfT/Vv+v33nuvGQFLA3n9ZlIbO/RBb0wVV4CW9HJ473VYRe3p/f3335u8YH1gDMFPObyR5eyUdSjFuXPnmnHNx48fb1JZZsyYYf6Y6AOyXOlX2iI0YMAAk5+u3+zoxIfHcnazOV0EELBNQEdhq1Wrlnnw1bXXXivTp083deu3kJp/vnz5ctMJVIej1dZ0TVfVtFUdpU3fT5kqrgBBesW991w5Ah4L6BPvdEjFN998U/bs2SPaEVSDdh2vV1vIX3rpJXee+bfffmty04t2dvL4QBREAAEE/EhA3yd1hLW1a9dK165dzdOUtWFNJx3b/J577jFDKeq35Do0rTZsaCdmVyDvRxRcymUIME76ZaCxCwL+JqApU7t27broZen45pqyUq9ePfnPf/5jOoZqK7qmuOhwddppWceV1klHHiBAvyglGxBAoAIJaPCtKYJpaWlmCFptxHBNuk072euIQjpalgbyq1evJkB3AfFbCNJ5ESCAgAm8NQjXVCr9mTZtmukMqk9k1AC+b9++0qpVKzOOtI5AoHnn+hWtplxp65B+LatDgDIhgAACFVlA88517HJ979Spbdu2Jt1PB3fQoZNTUlLk5Zf/v71zCaWtD8P4O/hiYOKuSEgoSslEkXYRkmKyKZcwVUQp8U0lMtoSIpekhJRyichEMXKJlAElZYJcBoSk7zxv7X04Ts52an0t+//8S3tba1nW/7e0vf3X8z7Pv7oPWRFY3ECBjoE+HhTtHCTgJsDGUTcJvpKAwQSio6O14D49PZWpqSmNPc/PzxdozSFvgQsBmpywIoR+CDySRVgR0vEQUAR/afZFGPwHxKmTgGEE8NkHKQscgd4OBA+hUIdEEHIWLGLMzMxoQyisa29ubmRxcVHGxsbUjx/ZJ1gE4SCB3xGgJv13VLiNBAwkgAYmNCkhyW58fFwdB1C0l5WVSX9/v1qC7e3tqW4S8hYkiKalpXl8+w1EximTAAkYRmB9fV1tZLe2tqSvr08drvCZ+evAAgfSQRFShLTQp6cnfQKJ4xDsdnh4KNnZ2Vw5/xUcv39HgEX6Oxz8hgTMJoAADUhYUKS7B5pCIyIi1K0Fj2mhn6ypqaHrgBsQX0mABIwhgFVvWCZiJR2JyejLwUo5BpJD4cyCgnxgYEAgZ2ltbdV9eNIIiSAHCXyFADXpX6HFY0nAxwkglAirPJubm56ZQs7y8PCgK+ZYGVpeXmaB7qHDNyRAAiYRQAYEZCwo0n/kzMjZ2Zl+Jh4cHKhzC/zNkcSM4XA4ZGJiQqUuKOA5SOCrBFikf5UYjycBHyaA1Z76+nptbELaHYKIJicnJS8vz4dnzamRAAmQgHcEIPHDl8vl0mIcYUSwWExKStLE5dXVVW26d58tPj5e7WtfX191scO9na8k4A0Byl28ocRjSMAgAtBPVlRUqI4S/4AKCwvVt9cgBJwqCZAACXgIoNkTQW5IAoXkD45X+FwsLi5W6R+8zp1Op9rRen7ozRtIBEtLS7VhNDQ09M0eviWBzwmwSP+cD/eSgJEE9vf31cklPT3dyPlz0iRAAiQAAnBqqa2tlaCgILm/v1c7RTxhRPM8JC/z8/OqP4djC6SAgYGBH8AVFBSoFWNHR8eHfdxAAp8RYJH+GR3uIwESIAESIAESMJZASUmJNs6jHwepy2iuh7MVQtxyc3MlISFB3a9gtZiRkSHd3d0fWL28vKgH+ocd3EACfyDAIv0PgLibBEiABEiABEjATAIoyGFNe3V1pa5XsFs8Pj6WpaUlQXpoe3u7IH3Zz89PoD8PCAgwExRnbQkBNo5agpUnJQESIAESIAES+O4Epqen1XIRmnRIXGCtiDA3BBPd3t7qqnpkZKQmMrNA/+53237XzyLdfveEV0QCJEACJEACJPA/EYCfObTl7gEnlp6eHm0OxTZoz+HekpmZKYODg5qyDCesqKgoaWlpERTpHCRgBYF/rDgpz0kCJEACJEACJEACdiIwOzsrSFZubm7Wy4LfeWVlpYYPwdUKicoIH4JzC3zNoTNHwY7UUFjRFhUVCZxehoaGJDk52U5T47X4KAEW6T56YzktEiABEiABEiCBnwQuLy81Udm9pampSeLi4mR4eFjtFUNCQrTBE5aKSBONjY3VQxsbGyUsLExOTk6kurpaYmJi3KfgKwlYSoCNo5bi5clJgARIgARIgATsQmB3d1fm5uYEyaHQlaNwT01NFSSGXl9fq5RlZGREE5btcs28DnMJUJNu7r3nzEmABEiABEjAKAJ3d3fqZw6vc6yk+/v7y9HRkaaItrW1aaEOv3MOErADAcpd7HAXeA0kQAIkQAIkQAKWE3A4HOpt3tnZKWtra7qqjsCixMREOT8/V+eW5+dny6+Dv4AEvCFAuYs3lHgMCZAACZAACZDAtyaAJGU0gNbV1YnT6ZTy8nJBqjKCicLDw+Xi4kKbRHt7e7/1PHnxvkOAK+m+cy85ExIgARIgARIwngA8zUdHR5UDEkJTUlKkoaFBtre3Bc2hwcHBkpOTowU7dOkoyhcWFgSr7PjiIAG7EOBKul3uBK+DBEiABEiABEjgrwlAtrKzsyMul0vg0LKxsaGWiSsrK1qkowCvqqrS8z8+PmqhnpWVJV1dXX/9O/mDJGAlgf8A3P5oPq7ltI8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {
      "image/png": {
       "width": 500
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(Image('dependency-accuracy.png', width=500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Bahdanau\n",
    "\n",
    "```text\n",
    "               precision    recall  f1-score   support\n",
    "\n",
    "          PAD     1.0000    1.0000    1.0000    843055\n",
    "          acl     0.9406    0.9296    0.9351      2983\n",
    "        advcl     0.8924    0.8613    0.8766      1175\n",
    "       advmod     0.9549    0.9482    0.9515      4712\n",
    "         amod     0.9296    0.9100    0.9197      4135\n",
    "        appos     0.9312    0.9570    0.9439      2488\n",
    "          aux     1.0000    1.0000    1.0000         5\n",
    "         case     0.9809    0.9823    0.9816     10557\n",
    "           cc     0.9676    0.9795    0.9735      3170\n",
    "        ccomp     0.8598    0.8045    0.8312       404\n",
    "     compound     0.9201    0.9464    0.9331      6605\n",
    "compound:plur     0.9597    0.9630    0.9613       594\n",
    "         conj     0.9600    0.9572    0.9586      4158\n",
    "          cop     0.9670    0.9720    0.9695       966\n",
    "        csubj     0.8929    0.8333    0.8621        30\n",
    "   csubj:pass     0.8000    0.6667    0.7273        12\n",
    "          dep     0.8189    0.9259    0.8691       459\n",
    "          det     0.9558    0.9369    0.9463      4041\n",
    "        fixed     0.9337    0.8953    0.9141       535\n",
    "         flat     0.9724    0.9714    0.9719     10479\n",
    "         iobj     0.9286    0.7222    0.8125        18\n",
    "         mark     0.9210    0.9491    0.9349      1376\n",
    "         nmod     0.9355    0.9324    0.9340      3921\n",
    "        nsubj     0.9430    0.9538    0.9484      6345\n",
    "   nsubj:pass     0.9458    0.9053    0.9251      1985\n",
    "       nummod     0.9762    0.9787    0.9775      3854\n",
    "          obj     0.9495    0.9465    0.9480      5162\n",
    "          obl     0.9458    0.9543    0.9500      5599\n",
    "    parataxis     0.9268    0.8283    0.8748       367\n",
    "        punct     0.9978    0.9968    0.9973     16549\n",
    "         root     0.9743    0.9643    0.9693      5037\n",
    "        xcomp     0.8878    0.9039    0.8958      1217\n",
    "\n",
    "  avg / total     0.9953    0.9953    0.9953    951993\n",
    "  \n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0     1.0000    1.0000    1.0000    843055\n",
    "          1     0.9718    0.9633    0.9675      5037\n",
    "          2     0.9604    0.9459    0.9531      4285\n",
    "          3     0.9474    0.9557    0.9515      4971\n",
    "          4     0.9575    0.9647    0.9611      6594\n",
    "          5     0.9534    0.9665    0.9599      5880\n",
    "          6     0.9648    0.9632    0.9640      6037\n",
    "          7     0.9512    0.9654    0.9582      5548\n",
    "          8     0.9611    0.9623    0.9617      5542\n",
    "          9     0.9729    0.9498    0.9612      4877\n",
    "         10     0.9614    0.9621    0.9617      4559\n",
    "         11     0.9495    0.9588    0.9541      4316\n",
    "         12     0.9547    0.9573    0.9560      3698\n",
    "         13     0.9664    0.9506    0.9584      3600\n",
    "         14     0.9652    0.9590    0.9621      3294\n",
    "         15     0.9619    0.9541    0.9580      3179\n",
    "         16     0.9604    0.9573    0.9589      3117\n",
    "         17     0.9634    0.9587    0.9610      2831\n",
    "         18     0.9406    0.9594    0.9499      2392\n",
    "         19     0.9657    0.9582    0.9619      2176\n",
    "         20     0.9656    0.9615    0.9635      2102\n",
    "         21     0.9523    0.9577    0.9550      1960\n",
    "         22     0.9519    0.9586    0.9552      1859\n",
    "         23     0.9605    0.9555    0.9580      1732\n",
    "         24     0.9649    0.9474    0.9561      1540\n",
    "         25     0.9399    0.9503    0.9451      1349\n",
    "         26     0.9680    0.9333    0.9503      1199\n",
    "         27     0.9246    0.9604    0.9422      1111\n",
    "         28     0.9491    0.9561    0.9526       956\n",
    "         29     0.9578    0.9646    0.9612       989\n",
    "         30     0.9365    0.9513    0.9438      1007\n",
    "         31     0.9483    0.9592    0.9537       784\n",
    "         32     0.9352    0.9545    0.9448       726\n",
    "         33     0.9468    0.9290    0.9378       690\n",
    "         34     0.9575    0.9464    0.9519       690\n",
    "         35     0.9480    0.9231    0.9354       533\n",
    "         36     0.9532    0.9432    0.9481       475\n",
    "         37     0.9511    0.9340    0.9425       500\n",
    "         38     0.9455    0.9139    0.9294       418\n",
    "         39     0.9326    0.9708    0.9513       342\n",
    "         40     0.9361    0.9338    0.9350       408\n",
    "         41     0.9260    0.9602    0.9428       352\n",
    "         42     0.9649    0.9615    0.9632       286\n",
    "         43     0.9418    0.9487    0.9453       273\n",
    "         44     0.9125    0.9389    0.9255       311\n",
    "         45     0.9406    0.9556    0.9480       315\n",
    "         46     0.9703    0.9655    0.9679       203\n",
    "         47     0.9662    0.9542    0.9602       240\n",
    "         48     0.9065    0.9065    0.9065       214\n",
    "         49     0.9455    0.9720    0.9585       214\n",
    "         50     0.9315    0.9189    0.9252       148\n",
    "         51     0.9356    0.9265    0.9310       204\n",
    "         52     0.9257    0.9580    0.9416       143\n",
    "         53     0.9496    0.9231    0.9362       143\n",
    "         54     0.9381    0.9430    0.9406       193\n",
    "         55     0.9837    0.9237    0.9528       131\n",
    "         56     0.8532    0.9688    0.9073        96\n",
    "         57     0.9604    0.9510    0.9557       102\n",
    "         58     0.9633    0.9459    0.9545       111\n",
    "         59     0.9870    0.8837    0.9325        86\n",
    "         60     1.0000    0.9559    0.9774        68\n",
    "         61     0.9429    0.9519    0.9474       104\n",
    "         62     0.9726    0.8875    0.9281        80\n",
    "         63     0.9459    0.9589    0.9524        73\n",
    "         64     0.9385    0.9531    0.9457        64\n",
    "         65     1.0000    0.8833    0.9381        60\n",
    "         66     0.8676    0.9516    0.9077        62\n",
    "         67     0.9020    0.8519    0.8762        54\n",
    "         68     0.9683    0.9242    0.9457        66\n",
    "         69     0.9474    0.9351    0.9412        77\n",
    "         70     0.8406    0.8923    0.8657        65\n",
    "         71     0.9474    0.9818    0.9643        55\n",
    "         72     0.9722    0.9459    0.9589        37\n",
    "         73     0.9796    0.9600    0.9697        50\n",
    "         74     0.9630    0.9630    0.9630        27\n",
    "         75     0.9750    1.0000    0.9873        39\n",
    "         76     0.9655    1.0000    0.9825        28\n",
    "         77     0.9655    0.9333    0.9492        30\n",
    "         78     1.0000    1.0000    1.0000        24\n",
    "         79     0.9677    1.0000    0.9836        30\n",
    "         80     0.9608    0.9074    0.9333        54\n",
    "         81     0.9167    1.0000    0.9565        11\n",
    "         82     0.9074    0.9423    0.9245        52\n",
    "         83     0.9259    1.0000    0.9615        25\n",
    "         84     0.9677    1.0000    0.9836        30\n",
    "         85     1.0000    1.0000    1.0000        14\n",
    "         86     1.0000    0.9412    0.9697        34\n",
    "         87     1.0000    1.0000    1.0000        22\n",
    "         88     1.0000    1.0000    1.0000         8\n",
    "         89     1.0000    1.0000    1.0000        14\n",
    "         90     1.0000    1.0000    1.0000        18\n",
    "         91     0.9677    0.8824    0.9231        34\n",
    "         92     0.8182    1.0000    0.9000         9\n",
    "         93     0.9444    0.9444    0.9444        18\n",
    "         94     1.0000    0.9444    0.9714        18\n",
    "         95     0.9259    0.9615    0.9434        26\n",
    "         96     1.0000    1.0000    1.0000         8\n",
    "         97     1.0000    1.0000    1.0000         2\n",
    "         98     1.0000    1.0000    1.0000        16\n",
    "         99     0.9697    0.8649    0.9143        37\n",
    "        100     1.0000    1.0000    1.0000         2\n",
    "        101     1.0000    1.0000    1.0000        44\n",
    "        102     1.0000    1.0000    1.0000        15\n",
    "        103     0.8889    1.0000    0.9412         8\n",
    "        104     0.8269    0.9773    0.8958        44\n",
    "        105     1.0000    1.0000    1.0000         6\n",
    "        106     1.0000    1.0000    1.0000         7\n",
    "        107     1.0000    1.0000    1.0000        10\n",
    "        108     0.9412    1.0000    0.9697        32\n",
    "        109     1.0000    1.0000    1.0000        13\n",
    "        110     1.0000    1.0000    1.0000         9\n",
    "        111     1.0000    1.0000    1.0000         1\n",
    "        112     1.0000    0.7826    0.8780        23\n",
    "        113     1.0000    1.0000    1.0000        16\n",
    "        114     0.8333    1.0000    0.9091         5\n",
    "        115     1.0000    1.0000    1.0000         1\n",
    "        116     0.9130    0.9545    0.9333        22\n",
    "        117     1.0000    1.0000    1.0000         5\n",
    "        118     0.0000    0.0000    0.0000         0\n",
    "        119     1.0000    1.0000    1.0000         3\n",
    "        120     1.0000    1.0000    1.0000        15\n",
    "        122     1.0000    1.0000    1.0000         8\n",
    "        123     1.0000    1.0000    1.0000         4\n",
    "        125     1.0000    1.0000    1.0000        10\n",
    "        126     1.0000    1.0000    1.0000         2\n",
    "        129     1.0000    1.0000    1.0000         8\n",
    "        133     1.0000    1.0000    1.0000         4\n",
    "        135     1.0000    1.0000    1.0000         3\n",
    "        136     1.0000    1.0000    1.0000         2\n",
    "        139     1.0000    1.0000    1.0000         1\n",
    "        142     1.0000    1.0000    1.0000         2\n",
    "        146     1.0000    1.0000    1.0000         1\n",
    "        151     1.0000    1.0000    1.0000         1\n",
    "\n",
    "avg / total     0.9951    0.9951    0.9951    951993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Luong\n",
    "\n",
    "```text\n",
    "               precision    recall  f1-score   support\n",
    "\n",
    "          PAD     1.0000    1.0000    1.0000    840905\n",
    "          acl     0.9249    0.9392    0.9320      3094\n",
    "        advcl     0.8952    0.8478    0.8709      1209\n",
    "       advmod     0.9629    0.9475    0.9551      4952\n",
    "         amod     0.9288    0.9246    0.9267      4218\n",
    "        appos     0.9535    0.9204    0.9367      2426\n",
    "          aux     1.0000    1.0000    1.0000         1\n",
    "         case     0.9796    0.9795    0.9796     10991\n",
    "           cc     0.9686    0.9739    0.9713      3298\n",
    "        ccomp     0.8426    0.8501    0.8463       447\n",
    "     compound     0.9170    0.9477    0.9321      6787\n",
    "compound:plur     0.9428    0.9744    0.9584       626\n",
    "         conj     0.9539    0.9581    0.9560      4251\n",
    "          cop     0.9625    0.9809    0.9716       993\n",
    "        csubj     0.9655    0.8750    0.9180        32\n",
    "   csubj:pass     1.0000    0.9167    0.9565        12\n",
    "          dep     0.8905    0.8320    0.8603       518\n",
    "          det     0.9503    0.9364    0.9433      4088\n",
    "        fixed     0.9113    0.8899    0.9005       554\n",
    "         flat     0.9596    0.9792    0.9693     10272\n",
    "         iobj     1.0000    0.6000    0.7500        15\n",
    "         mark     0.9396    0.9217    0.9305      1417\n",
    "         nmod     0.9086    0.9475    0.9277      4155\n",
    "        nsubj     0.9524    0.9547    0.9535      6483\n",
    "   nsubj:pass     0.9402    0.9108    0.9252      1916\n",
    "       nummod     0.9747    0.9761    0.9754      4022\n",
    "          obj     0.9559    0.9468    0.9513      5337\n",
    "          obl     0.9622    0.9242    0.9428      5727\n",
    "    parataxis     0.8072    0.8910    0.8470       376\n",
    "        punct     0.9972    0.9984    0.9978     16581\n",
    "         root     0.9646    0.9688    0.9667      5037\n",
    "        xcomp     0.9225    0.8364    0.8774      1253\n",
    "\n",
    "  avg / total     0.9950    0.9950    0.9950    951993\n",
    "  \n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0     1.0000    1.0000    1.0000    840905\n",
    "          1     0.9709    0.9726    0.9717      5037\n",
    "          2     0.9310    0.9534    0.9420      4271\n",
    "          3     0.9543    0.9485    0.9514      5148\n",
    "          4     0.9587    0.9514    0.9551      6220\n",
    "          5     0.9471    0.9631    0.9550      5984\n",
    "          6     0.9593    0.9585    0.9589      5827\n",
    "          7     0.9597    0.9554    0.9576      5789\n",
    "          8     0.9657    0.9527    0.9592      5559\n",
    "          9     0.9548    0.9517    0.9532      5088\n",
    "         10     0.9565    0.9492    0.9528      4427\n",
    "         11     0.9458    0.9631    0.9544      4280\n",
    "         12     0.9584    0.9540    0.9562      3910\n",
    "         13     0.9481    0.9586    0.9533      3791\n",
    "         14     0.9385    0.9563    0.9473      3272\n",
    "         15     0.9577    0.9389    0.9482      3306\n",
    "         16     0.9383    0.9560    0.9471      3023\n",
    "         17     0.9629    0.9417    0.9522      2815\n",
    "         18     0.9384    0.9548    0.9465      2409\n",
    "         19     0.9463    0.9391    0.9427      2103\n",
    "         20     0.9349    0.9617    0.9481      2166\n",
    "         21     0.9712    0.9354    0.9530      2090\n",
    "         22     0.9525    0.9450    0.9487      1763\n",
    "         23     0.9512    0.9512    0.9512      1742\n",
    "         24     0.9624    0.9475    0.9549      1619\n",
    "         25     0.9439    0.9460    0.9449      1333\n",
    "         26     0.9584    0.9333    0.9457      1260\n",
    "         27     0.9443    0.9231    0.9336      1158\n",
    "         28     0.9384    0.9414    0.9399       955\n",
    "         29     0.9313    0.9417    0.9365      1080\n",
    "         30     0.9332    0.9323    0.9327      1004\n",
    "         31     0.9240    0.9404    0.9322       789\n",
    "         32     0.9500    0.9226    0.9361       762\n",
    "         33     0.9292    0.9502    0.9396       843\n",
    "         34     0.9553    0.9468    0.9510       677\n",
    "         35     0.9284    0.9396    0.9339       662\n",
    "         36     0.9238    0.9287    0.9262       561\n",
    "         37     0.9213    0.9152    0.9183       448\n",
    "         38     0.8978    0.9114    0.9045       395\n",
    "         39     0.8991    0.9114    0.9052       440\n",
    "         40     0.9262    0.9446    0.9353       505\n",
    "         41     0.9289    0.9098    0.9193       388\n",
    "         42     0.9544    0.9181    0.9359       342\n",
    "         43     0.9119    0.9308    0.9212       289\n",
    "         44     0.9106    0.9006    0.9056       362\n",
    "         45     0.8525    0.9091    0.8799       286\n",
    "         46     0.9283    0.8859    0.9066       263\n",
    "         47     0.9068    0.8924    0.8995       316\n",
    "         48     0.9282    0.9095    0.9188       199\n",
    "         49     0.9648    0.9202    0.9419       238\n",
    "         50     0.9274    0.9583    0.9426       120\n",
    "         51     0.9167    0.9585    0.9371       241\n",
    "         52     0.9507    0.9415    0.9461       205\n",
    "         53     0.9248    0.9179    0.9213       134\n",
    "         54     0.9200    0.9306    0.9253       173\n",
    "         55     0.9329    0.8910    0.9115       156\n",
    "         56     0.9073    0.8954    0.9013       153\n",
    "         57     0.9304    0.9469    0.9386       113\n",
    "         58     0.9417    0.9576    0.9496       118\n",
    "         59     0.8947    0.8500    0.8718       100\n",
    "         60     0.9770    0.8095    0.8854       105\n",
    "         61     0.8020    0.9576    0.8729       165\n",
    "         62     0.8767    0.8889    0.8828        72\n",
    "         63     0.9355    0.8365    0.8832       104\n",
    "         64     0.8852    0.8308    0.8571        65\n",
    "         65     0.9375    0.8955    0.9160        67\n",
    "         66     0.8690    0.8588    0.8639        85\n",
    "         67     0.9839    0.8472    0.9104        72\n",
    "         68     0.9223    0.9500    0.9360       100\n",
    "         69     0.9367    0.9250    0.9308        80\n",
    "         70     0.8442    0.9701    0.9028        67\n",
    "         71     0.8462    0.8462    0.8462        65\n",
    "         72     0.9200    0.8734    0.8961        79\n",
    "         73     0.8909    0.8596    0.8750        57\n",
    "         74     0.9487    0.8810    0.9136        42\n",
    "         75     0.9296    0.8919    0.9103        74\n",
    "         76     0.8333    0.9677    0.8955        31\n",
    "         77     0.8056    0.9062    0.8529        32\n",
    "         78     0.8750    0.8077    0.8400        26\n",
    "         79     0.7636    0.9333    0.8400        45\n",
    "         80     0.9180    0.8889    0.9032        63\n",
    "         81     0.7188    0.8214    0.7667        28\n",
    "         82     0.8983    0.9298    0.9138        57\n",
    "         83     1.0000    0.8571    0.9231        28\n",
    "         84     0.8605    0.9487    0.9024        39\n",
    "         85     0.9474    0.9474    0.9474        19\n",
    "         86     0.8919    0.9706    0.9296        34\n",
    "         87     0.9231    0.8571    0.8889        14\n",
    "         88     0.9474    0.7826    0.8571        23\n",
    "         89     1.0000    0.8571    0.9231        14\n",
    "         90     0.8929    0.8621    0.8772        29\n",
    "         91     0.8462    0.9429    0.8919        35\n",
    "         92     0.9333    0.7568    0.8358        37\n",
    "         93     0.7895    0.8333    0.8108        18\n",
    "         94     1.0000    0.8000    0.8889        20\n",
    "         95     0.9048    0.9500    0.9268        20\n",
    "         96     0.9412    0.9412    0.9412        17\n",
    "         97     0.9583    1.0000    0.9787        23\n",
    "         98     0.9000    1.0000    0.9474         9\n",
    "         99     1.0000    0.9643    0.9818        28\n",
    "        100     0.8333    1.0000    0.9091         5\n",
    "        101     1.0000    0.9231    0.9600        13\n",
    "        102     1.0000    1.0000    1.0000        13\n",
    "        103     0.8750    1.0000    0.9333        14\n",
    "        104     1.0000    0.9231    0.9600        26\n",
    "        105     1.0000    0.9167    0.9565        12\n",
    "        106     0.9444    0.8500    0.8947        20\n",
    "        107     1.0000    0.8571    0.9231        21\n",
    "        108     1.0000    1.0000    1.0000        20\n",
    "        109     1.0000    1.0000    1.0000         6\n",
    "        110     0.8750    1.0000    0.9333         7\n",
    "        111     1.0000    1.0000    1.0000         4\n",
    "        112     0.9200    0.9583    0.9388        24\n",
    "        113     0.8889    1.0000    0.9412         8\n",
    "        114     1.0000    0.6667    0.8000         3\n",
    "        115     1.0000    1.0000    1.0000         5\n",
    "        116     0.9474    0.8571    0.9000        21\n",
    "        117     0.6667    1.0000    0.8000         2\n",
    "        119     1.0000    1.0000    1.0000         3\n",
    "        120     0.8824    0.9375    0.9091        16\n",
    "        121     1.0000    0.8000    0.8889         5\n",
    "        122     0.8889    1.0000    0.9412         8\n",
    "        123     0.0000    0.0000    0.0000         2\n",
    "        124     1.0000    0.6667    0.8000         3\n",
    "        125     1.0000    1.0000    1.0000         8\n",
    "        126     1.0000    0.8000    0.8889        10\n",
    "        127     1.0000    1.0000    1.0000         3\n",
    "        128     1.0000    1.0000    1.0000         1\n",
    "        129     1.0000    1.0000    1.0000         5\n",
    "        130     1.0000    0.8333    0.9091        12\n",
    "        131     1.0000    1.0000    1.0000         2\n",
    "        132     1.0000    1.0000    1.0000         1\n",
    "        133     1.0000    1.0000    1.0000         9\n",
    "        134     1.0000    1.0000    1.0000         6\n",
    "        136     1.0000    1.0000    1.0000         3\n",
    "        137     1.0000    1.0000    1.0000        10\n",
    "        138     1.0000    1.0000    1.0000        10\n",
    "        140     1.0000    1.0000    1.0000         4\n",
    "        141     1.0000    1.0000    1.0000         2\n",
    "        142     0.4000    0.5000    0.4444         4\n",
    "        144     0.5714    1.0000    0.7273         4\n",
    "        146     0.7500    0.7500    0.7500         4\n",
    "        147     1.0000    0.6000    0.7500         5\n",
    "        149     1.0000    1.0000    1.0000         2\n",
    "        150     0.6667    0.6667    0.6667         3\n",
    "        151     0.5000    0.5000    0.5000         2\n",
    "        152     0.5000    0.5000    0.5000         2\n",
    "        153     1.0000    0.5000    0.6667         2\n",
    "        156     1.0000    1.0000    1.0000         2\n",
    "        158     0.8889    1.0000    0.9412         8\n",
    "        160     0.8000    1.0000    0.8889         4\n",
    "        164     1.0000    1.0000    1.0000         4\n",
    "\n",
    "avg / total     0.9941    0.9941    0.9941    951993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Concat\n",
    "\n",
    "```text\n",
    "               precision    recall  f1-score   support\n",
    "\n",
    "          PAD     1.0000    1.0000    1.0000    841717\n",
    "          acl     0.9501    0.9110    0.9301      2965\n",
    "        advcl     0.8127    0.8719    0.8413      1249\n",
    "       advmod     0.9423    0.9329    0.9376      4846\n",
    "         amod     0.9141    0.9104    0.9123      4208\n",
    "        appos     0.9282    0.9266    0.9274      2412\n",
    "         case     0.9757    0.9756    0.9756     10896\n",
    "           cc     0.9613    0.9726    0.9669      3171\n",
    "        ccomp     0.8115    0.7094    0.7570       437\n",
    "     compound     0.9176    0.9350    0.9263      6804\n",
    "compound:plur     0.9172    0.9767    0.9460       601\n",
    "         conj     0.9504    0.9493    0.9498      4119\n",
    "          cop     0.9621    0.9761    0.9690       962\n",
    "        csubj     0.8095    0.7083    0.7556        24\n",
    "   csubj:pass     0.7500    0.6000    0.6667        10\n",
    "          dep     0.8712    0.8333    0.8519       552\n",
    "          det     0.9288    0.9339    0.9313      4082\n",
    "        fixed     0.9229    0.8288    0.8733       549\n",
    "         flat     0.9619    0.9712    0.9666     10328\n",
    "         iobj     0.7273    0.8000    0.7619        10\n",
    "         mark     0.9059    0.9260    0.9159      1487\n",
    "         nmod     0.9159    0.9318    0.9238      4105\n",
    "        nsubj     0.9284    0.9550    0.9415      6316\n",
    "   nsubj:pass     0.9367    0.8999    0.9179      2007\n",
    "       nummod     0.9743    0.9617    0.9680      4024\n",
    "          obj     0.9428    0.9340    0.9384      5184\n",
    "          obl     0.9598    0.9292    0.9442      5776\n",
    "    parataxis     0.8301    0.7537    0.7900       337\n",
    "        punct     0.9957    0.9984    0.9971     16529\n",
    "         root     0.9654    0.9694    0.9674      5037\n",
    "        xcomp     0.8955    0.8575    0.8761      1249\n",
    "\n",
    "  avg / total     0.9943    0.9943    0.9943    951993\n",
    "\n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0     1.0000    1.0000    1.0000    841717\n",
    "          1     0.9638    0.9676    0.9657      5037\n",
    "          2     0.9526    0.9295    0.9409      4367\n",
    "          3     0.9410    0.9395    0.9403      4942\n",
    "          4     0.9544    0.9516    0.9530      6440\n",
    "          5     0.9453    0.9514    0.9484      6035\n",
    "          6     0.9376    0.9633    0.9503      6024\n",
    "          7     0.9456    0.9491    0.9473      5398\n",
    "          8     0.9506    0.9438    0.9472      5482\n",
    "          9     0.9488    0.9455    0.9472      4977\n",
    "         10     0.9331    0.9578    0.9453      4430\n",
    "         11     0.9453    0.9468    0.9460      4583\n",
    "         12     0.9364    0.9420    0.9392      3673\n",
    "         13     0.9495    0.9298    0.9395      3719\n",
    "         14     0.9425    0.9343    0.9384      3316\n",
    "         15     0.9460    0.9197    0.9327      3065\n",
    "         16     0.9125    0.9443    0.9281      3071\n",
    "         17     0.9350    0.9228    0.9289      2667\n",
    "         18     0.9377    0.9198    0.9286      2469\n",
    "         19     0.9167    0.9267    0.9217      2197\n",
    "         20     0.9076    0.9286    0.9180      2031\n",
    "         21     0.9355    0.8701    0.9016      1917\n",
    "         22     0.8985    0.8980    0.8983      1834\n",
    "         23     0.9038    0.9011    0.9025      1689\n",
    "         24     0.9066    0.8968    0.9017      1667\n",
    "         25     0.8782    0.9227    0.8999      1320\n",
    "         26     0.8769    0.9204    0.8982      1169\n",
    "         27     0.9041    0.9049    0.9045      1094\n",
    "         28     0.9054    0.8825    0.8938       987\n",
    "         29     0.9352    0.8799    0.9067      1099\n",
    "         30     0.8952    0.9110    0.9031       910\n",
    "         31     0.8745    0.8951    0.8847       810\n",
    "         32     0.8978    0.8772    0.8874       741\n",
    "         33     0.8782    0.9206    0.8989       705\n",
    "         34     0.9467    0.8692    0.9063       818\n",
    "         35     0.8893    0.8745    0.8819       542\n",
    "         36     0.9258    0.8794    0.9020       539\n",
    "         37     0.8603    0.9259    0.8919       459\n",
    "         38     0.9019    0.8458    0.8729       402\n",
    "         39     0.8577    0.9035    0.8800       487\n",
    "         40     0.8374    0.9071    0.8709       420\n",
    "         41     0.9148    0.8496    0.8810       379\n",
    "         42     0.8424    0.9393    0.8882       313\n",
    "         43     0.8852    0.8415    0.8628       284\n",
    "         44     0.9130    0.8571    0.8842       245\n",
    "         45     0.8829    0.9009    0.8918       343\n",
    "         46     0.8036    0.8654    0.8333       208\n",
    "         47     0.8803    0.8834    0.8818       283\n",
    "         48     0.9158    0.7699    0.8365       226\n",
    "         49     0.9074    0.8376    0.8711       234\n",
    "         50     0.7014    0.9136    0.7936       162\n",
    "         51     0.8268    0.9080    0.8655       163\n",
    "         52     0.8539    0.8889    0.8711       171\n",
    "         53     0.9136    0.8457    0.8783       175\n",
    "         54     0.8881    0.8581    0.8729       148\n",
    "         55     0.9073    0.8354    0.8698       164\n",
    "         56     0.8456    0.9200    0.8812       125\n",
    "         57     0.9000    0.8250    0.8609       120\n",
    "         58     0.9027    0.8430    0.8718       121\n",
    "         59     0.7947    0.9231    0.8541       130\n",
    "         60     0.7705    0.7833    0.7769        60\n",
    "         61     0.9315    0.8774    0.9037       155\n",
    "         62     0.8611    0.8493    0.8552        73\n",
    "         63     0.8172    0.9048    0.8588        84\n",
    "         64     0.8571    0.7273    0.7869        66\n",
    "         65     0.9130    0.8750    0.8936        72\n",
    "         66     0.7500    0.9398    0.8342        83\n",
    "         67     0.8409    0.8315    0.8362        89\n",
    "         68     0.9545    0.7590    0.8456        83\n",
    "         69     0.8916    0.8810    0.8862        84\n",
    "         70     0.7727    0.8644    0.8160        59\n",
    "         71     0.8679    0.8846    0.8762        52\n",
    "         72     0.8876    0.8404    0.8634        94\n",
    "         73     0.9298    0.8833    0.9060        60\n",
    "         74     0.9273    0.8226    0.8718        62\n",
    "         75     0.9070    0.8298    0.8667        47\n",
    "         76     0.7885    0.8723    0.8283        47\n",
    "         77     0.8000    0.8780    0.8372        41\n",
    "         78     0.8542    1.0000    0.9213        41\n",
    "         79     0.8696    0.9091    0.8889        44\n",
    "         80     0.9375    0.8571    0.8955        70\n",
    "         81     0.8667    0.7222    0.7879        36\n",
    "         82     0.8514    0.9130    0.8811        69\n",
    "         83     0.9024    0.9250    0.9136        40\n",
    "         84     0.9444    1.0000    0.9714        34\n",
    "         85     0.9189    0.9444    0.9315        36\n",
    "         86     0.8810    0.9487    0.9136        39\n",
    "         87     0.9310    0.8710    0.9000        31\n",
    "         88     0.8857    1.0000    0.9394        31\n",
    "         89     0.9200    0.9200    0.9200        25\n",
    "         90     0.8667    0.8125    0.8387        32\n",
    "         91     0.8519    0.9200    0.8846        25\n",
    "         92     0.8913    0.9535    0.9213        43\n",
    "         93     0.8500    0.9444    0.8947        18\n",
    "         94     0.9231    0.8571    0.8889        28\n",
    "         95     0.7500    0.8571    0.8000         7\n",
    "         96     0.9375    0.7143    0.8108        21\n",
    "         97     0.9688    0.8158    0.8857        38\n",
    "         98     0.9091    0.8696    0.8889        23\n",
    "         99     0.8462    1.0000    0.9167        33\n",
    "        100     1.0000    0.7778    0.8750         9\n",
    "        101     0.9744    0.9744    0.9744        39\n",
    "        102     0.8636    0.8636    0.8636        22\n",
    "        103     0.9677    0.9677    0.9677        31\n",
    "        104     1.0000    1.0000    1.0000         7\n",
    "        105     1.0000    0.6471    0.7857        17\n",
    "        106     0.9600    1.0000    0.9796        24\n",
    "        107     0.9750    1.0000    0.9873        39\n",
    "        108     0.8947    1.0000    0.9444        17\n",
    "        109     1.0000    1.0000    1.0000        14\n",
    "        110     0.9524    1.0000    0.9756        20\n",
    "        111     0.9091    0.8333    0.8696        12\n",
    "        112     0.9259    0.9259    0.9259        27\n",
    "        113     0.8889    1.0000    0.9412        16\n",
    "        114     0.8000    0.9231    0.8571        13\n",
    "        115     0.8235    1.0000    0.9032        14\n",
    "        116     1.0000    0.8095    0.8947        21\n",
    "        117     1.0000    0.8571    0.9231         7\n",
    "        118     0.7692    0.8333    0.8000        12\n",
    "        119     1.0000    1.0000    1.0000         4\n",
    "        120     0.9500    1.0000    0.9744        19\n",
    "        121     1.0000    1.0000    1.0000         7\n",
    "        122     0.8235    0.9333    0.8750        15\n",
    "        123     1.0000    1.0000    1.0000         6\n",
    "        124     1.0000    0.3333    0.5000         3\n",
    "        125     1.0000    0.8889    0.9412        18\n",
    "        126     1.0000    0.9667    0.9831        30\n",
    "        127     0.8750    1.0000    0.9333         7\n",
    "        128     0.8333    0.8333    0.8333         6\n",
    "        129     0.9412    0.9412    0.9412        17\n",
    "        130     0.9333    1.0000    0.9655        14\n",
    "        131     1.0000    1.0000    1.0000         9\n",
    "        132     1.0000    1.0000    1.0000         3\n",
    "        133     1.0000    1.0000    1.0000        11\n",
    "        134     0.9412    1.0000    0.9697        16\n",
    "        135     1.0000    1.0000    1.0000         6\n",
    "        136     1.0000    0.8000    0.8889        10\n",
    "        137     1.0000    0.8000    0.8889        10\n",
    "        138     1.0000    1.0000    1.0000        22\n",
    "        139     0.0000    0.0000    0.0000         1\n",
    "        140     1.0000    1.0000    1.0000         2\n",
    "        141     1.0000    1.0000    1.0000         2\n",
    "        142     1.0000    1.0000    1.0000         4\n",
    "        144     1.0000    1.0000    1.0000         4\n",
    "        146     1.0000    1.0000    1.0000         3\n",
    "        147     0.8889    1.0000    0.9412         8\n",
    "        149     1.0000    1.0000    1.0000         4\n",
    "        150     0.7500    1.0000    0.8571         3\n",
    "        151     1.0000    1.0000    1.0000         2\n",
    "        152     1.0000    1.0000    1.0000         1\n",
    "        153     1.0000    1.0000    1.0000         1\n",
    "        154     1.0000    1.0000    1.0000         2\n",
    "        156     1.0000    0.8333    0.9091         6\n",
    "        157     1.0000    1.0000    1.0000         1\n",
    "        158     1.0000    1.0000    1.0000         5\n",
    "        159     1.0000    1.0000    1.0000         1\n",
    "        160     1.0000    1.0000    1.0000         2\n",
    "        162     1.0000    1.0000    1.0000         3\n",
    "        163     0.0000    0.0000    0.0000         2\n",
    "        164     1.0000    1.0000    1.0000         2\n",
    "        167     0.6667    1.0000    0.8000         4\n",
    "        174     1.0000    1.0000    1.0000         2\n",
    "        176     1.0000    0.7500    0.8571         4\n",
    "        177     1.0000    1.0000    1.0000         2\n",
    "        178     1.0000    1.0000    1.0000         1\n",
    "        179     1.0000    1.0000    1.0000         1\n",
    "        182     1.0000    1.0000    1.0000         4\n",
    "        183     1.0000    1.0000    1.0000         4\n",
    "\n",
    "avg / total     0.9921    0.9920    0.9920    951993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Attention is all you need\n",
    "\n",
    "```text\n",
    "               precision    recall  f1-score   support\n",
    "\n",
    "          PAD     1.0000    1.0000    1.0000    841796\n",
    "          acl     0.8768    0.8849    0.8809      3016\n",
    "        advcl     0.8290    0.7943    0.8113      1196\n",
    "       advmod     0.9043    0.9163    0.9102      4754\n",
    "         amod     0.9121    0.8773    0.8943      4149\n",
    "        appos     0.8934    0.8983    0.8958      2547\n",
    "          aux     1.0000    1.0000    1.0000         6\n",
    "         case     0.9593    0.9670    0.9631     10888\n",
    "           cc     0.9523    0.9606    0.9564      3198\n",
    "        ccomp     0.7984    0.7385    0.7673       413\n",
    "     compound     0.8677    0.8956    0.8815      6679\n",
    "compound:plur     0.9073    0.9255    0.9163       550\n",
    "         conj     0.8625    0.9330    0.8964      4162\n",
    "          cop     0.9296    0.9679    0.9484       996\n",
    "        csubj     0.9000    0.4091    0.5625        22\n",
    "   csubj:pass     0.8462    0.8462    0.8462        13\n",
    "          dep     0.8274    0.7377    0.7800       507\n",
    "          det     0.8897    0.9196    0.9044      4094\n",
    "        fixed     0.8851    0.7966    0.8385       580\n",
    "         flat     0.9468    0.9198    0.9331     10333\n",
    "         iobj     1.0000    0.6000    0.7500        20\n",
    "         mark     0.8535    0.8447    0.8491      1359\n",
    "         nmod     0.8749    0.8907    0.8827      4107\n",
    "        nsubj     0.8746    0.8881    0.8813      6471\n",
    "   nsubj:pass     0.8478    0.7116    0.7738      1949\n",
    "       nummod     0.9568    0.9524    0.9546      3884\n",
    "          obj     0.9082    0.8946    0.9013      5274\n",
    "          obl     0.9203    0.8854    0.9025      5740\n",
    "    parataxis     0.7980    0.7980    0.7980       391\n",
    "        punct     0.9933    0.9957    0.9945     16561\n",
    "         root     0.8974    0.9200    0.9085      5037\n",
    "        xcomp     0.8580    0.8593    0.8587      1301\n",
    "\n",
    "  avg / total     0.9906    0.9906    0.9906    951993\n",
    "  \n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          0     1.0000    1.0000    1.0000    841796\n",
    "          1     0.9486    0.9277    0.9381      5037\n",
    "          2     0.9157    0.9547    0.9348      4325\n",
    "          3     0.9505    0.9137    0.9318      4856\n",
    "          4     0.9439    0.9311    0.9374      6309\n",
    "          5     0.9422    0.9396    0.9409      6540\n",
    "          6     0.9314    0.9516    0.9414      5697\n",
    "          7     0.9468    0.9461    0.9464      5414\n",
    "          8     0.9524    0.9394    0.9458      5559\n",
    "          9     0.9432    0.9421    0.9427      5028\n",
    "         10     0.9308    0.9544    0.9425      4300\n",
    "         11     0.9623    0.9323    0.9471      4358\n",
    "         12     0.9449    0.9493    0.9471      3903\n",
    "         13     0.9338    0.9442    0.9390      3497\n",
    "         14     0.9444    0.9475    0.9459      3445\n",
    "         15     0.9445    0.9487    0.9466      3177\n",
    "         16     0.9411    0.9589    0.9500      3068\n",
    "         17     0.9350    0.9589    0.9468      2774\n",
    "         18     0.9527    0.9352    0.9439      2499\n",
    "         19     0.9767    0.9207    0.9478      2319\n",
    "         20     0.9445    0.9558    0.9501      2013\n",
    "         21     0.9321    0.9374    0.9347      2124\n",
    "         22     0.9337    0.9423    0.9380      1749\n",
    "         23     0.9508    0.9175    0.9339      1685\n",
    "         24     0.9608    0.9240    0.9421      1540\n",
    "         25     0.8654    0.9661    0.9130      1358\n",
    "         26     0.9511    0.9245    0.9376      1179\n",
    "         27     0.9416    0.9367    0.9392      1154\n",
    "         28     0.8961    0.9549    0.9245       975\n",
    "         29     0.9260    0.9383    0.9321      1054\n",
    "         30     0.9342    0.9551    0.9445      1025\n",
    "         31     0.9482    0.9146    0.9311       761\n",
    "         32     0.9549    0.9126    0.9333       835\n",
    "         33     0.9235    0.9506    0.9368       749\n",
    "         34     0.9492    0.9465    0.9478       710\n",
    "         35     0.9323    0.9649    0.9483       599\n",
    "         36     0.9750    0.9458    0.9602       535\n",
    "         37     0.9363    0.9620    0.9490       474\n",
    "         38     0.9099    0.9815    0.9443       432\n",
    "         39     0.9462    0.9342    0.9401       395\n",
    "         40     0.9170    0.9535    0.9349       452\n",
    "         41     0.9446    0.9214    0.9328       407\n",
    "         42     0.9452    0.9452    0.9452       292\n",
    "         43     0.9731    0.9031    0.9368       320\n",
    "         44     0.9030    0.9767    0.9384       343\n",
    "         45     0.9343    0.9812    0.9572       319\n",
    "         46     0.9943    0.7955    0.8838       220\n",
    "         47     0.9420    0.9684    0.9550       285\n",
    "         48     0.9160    0.9745    0.9443       235\n",
    "         49     0.9113    0.9893    0.9487       187\n",
    "         50     0.9568    0.8636    0.9078       154\n",
    "         51     0.9706    0.9538    0.9621       173\n",
    "         52     0.9554    0.9934    0.9740       151\n",
    "         53     0.9116    0.9515    0.9311       206\n",
    "         54     0.9008    0.9833    0.9402       120\n",
    "         55     0.9371    0.9371    0.9371       159\n",
    "         56     0.9179    0.9535    0.9354       129\n",
    "         57     0.9091    0.8824    0.8955       102\n",
    "         58     0.9350    0.9127    0.9237       126\n",
    "         59     0.9725    0.7910    0.8724       134\n",
    "         60     0.9576    0.9826    0.9700       115\n",
    "         61     0.9200    0.9485    0.9340        97\n",
    "         62     0.9200    0.9079    0.9139        76\n",
    "         63     0.9551    0.9770    0.9659        87\n",
    "         64     0.9878    0.9310    0.9586        87\n",
    "         65     0.9103    0.9861    0.9467        72\n",
    "         66     0.9474    0.9863    0.9664        73\n",
    "         67     1.0000    0.9667    0.9831        60\n",
    "         68     0.9855    0.8831    0.9315        77\n",
    "         69     0.8889    0.9231    0.9057        52\n",
    "         70     0.9524    1.0000    0.9756        80\n",
    "         71     0.9241    0.9605    0.9419        76\n",
    "         72     0.9870    0.9870    0.9870        77\n",
    "         73     0.9531    1.0000    0.9760        61\n",
    "         74     1.0000    0.9667    0.9831        30\n",
    "         75     0.9412    1.0000    0.9697        64\n",
    "         76     1.0000    0.8571    0.9231        28\n",
    "         77     0.9487    1.0000    0.9737        37\n",
    "         78     0.9677    0.9677    0.9677        31\n",
    "         79     1.0000    1.0000    1.0000        25\n",
    "         80     1.0000    0.9348    0.9663        46\n",
    "         81     1.0000    0.9756    0.9877        41\n",
    "         82     1.0000    0.9302    0.9639        43\n",
    "         83     0.9474    1.0000    0.9730        18\n",
    "         84     0.8846    1.0000    0.9388        23\n",
    "         85     0.9583    1.0000    0.9787        23\n",
    "         86     1.0000    0.8636    0.9268        44\n",
    "         87     1.0000    1.0000    1.0000        10\n",
    "         88     0.9412    0.9412    0.9412        17\n",
    "         89     1.0000    0.8750    0.9333         8\n",
    "         90     0.9167    0.9565    0.9362        23\n",
    "         91     1.0000    1.0000    1.0000        15\n",
    "         92     1.0000    1.0000    1.0000        34\n",
    "         93     0.8571    1.0000    0.9231         6\n",
    "         94     0.9231    1.0000    0.9600        12\n",
    "         95     1.0000    1.0000    1.0000         9\n",
    "         96     1.0000    0.9333    0.9655        15\n",
    "         97     1.0000    1.0000    1.0000        30\n",
    "         98     1.0000    1.0000    1.0000         8\n",
    "         99     1.0000    0.9200    0.9583        25\n",
    "        100     0.8571    1.0000    0.9231         6\n",
    "        101     1.0000    0.9744    0.9870        39\n",
    "        102     1.0000    1.0000    1.0000         7\n",
    "        103     0.8889    1.0000    0.9412        16\n",
    "        104     1.0000    0.9500    0.9744        20\n",
    "        105     1.0000    0.9000    0.9474        10\n",
    "        106     0.9500    1.0000    0.9744        19\n",
    "        107     0.7500    1.0000    0.8571        27\n",
    "        108     1.0000    1.0000    1.0000        15\n",
    "        109     1.0000    1.0000    1.0000         3\n",
    "        110     1.0000    1.0000    1.0000        14\n",
    "        111     1.0000    1.0000    1.0000         9\n",
    "        112     0.9474    1.0000    0.9730        18\n",
    "        113     0.8571    1.0000    0.9231         6\n",
    "        114     1.0000    1.0000    1.0000        10\n",
    "        115     1.0000    1.0000    1.0000         7\n",
    "        116     1.0000    0.9375    0.9677        16\n",
    "        117     1.0000    0.5000    0.6667         2\n",
    "        118     1.0000    1.0000    1.0000        12\n",
    "        119     1.0000    1.0000    1.0000         4\n",
    "        120     1.0000    0.9231    0.9600        13\n",
    "        121     1.0000    1.0000    1.0000         6\n",
    "        122     1.0000    1.0000    1.0000         3\n",
    "        123     1.0000    0.8333    0.9091         6\n",
    "        124     1.0000    1.0000    1.0000         2\n",
    "        125     1.0000    1.0000    1.0000         2\n",
    "        126     0.8846    1.0000    0.9388        23\n",
    "        127     1.0000    1.0000    1.0000         6\n",
    "        128     1.0000    1.0000    1.0000         5\n",
    "        129     1.0000    0.8333    0.9091         6\n",
    "        130     1.0000    1.0000    1.0000        12\n",
    "        131     1.0000    0.7143    0.8333         7\n",
    "        132     1.0000    1.0000    1.0000         2\n",
    "        133     1.0000    1.0000    1.0000         4\n",
    "        134     0.9000    0.9000    0.9000        10\n",
    "        135     0.8571    1.0000    0.9231         6\n",
    "        136     1.0000    1.0000    1.0000         7\n",
    "        137     1.0000    1.0000    1.0000         8\n",
    "        138     1.0000    1.0000    1.0000        12\n",
    "        139     1.0000    1.0000    1.0000         1\n",
    "        140     1.0000    1.0000    1.0000         2\n",
    "        141     1.0000    1.0000    1.0000         2\n",
    "        142     1.0000    1.0000    1.0000         4\n",
    "        144     1.0000    1.0000    1.0000         4\n",
    "        146     1.0000    1.0000    1.0000         3\n",
    "        147     1.0000    1.0000    1.0000         7\n",
    "        149     1.0000    1.0000    1.0000         2\n",
    "        150     1.0000    1.0000    1.0000         2\n",
    "        151     1.0000    1.0000    1.0000         2\n",
    "        152     1.0000    1.0000    1.0000         1\n",
    "        153     1.0000    1.0000    1.0000         1\n",
    "        154     1.0000    1.0000    1.0000         2\n",
    "        156     1.0000    1.0000    1.0000         6\n",
    "        157     1.0000    1.0000    1.0000         1\n",
    "        158     1.0000    1.0000    1.0000         5\n",
    "        159     1.0000    1.0000    1.0000         1\n",
    "        160     1.0000    1.0000    1.0000         2\n",
    "        162     0.6667    0.6667    0.6667         3\n",
    "        163     0.6667    1.0000    0.8000         2\n",
    "        164     1.0000    1.0000    1.0000         2\n",
    "        167     1.0000    0.7500    0.8571         4\n",
    "        174     1.0000    1.0000    1.0000         2\n",
    "        176     1.0000    1.0000    1.0000         4\n",
    "        177     1.0000    1.0000    1.0000         2\n",
    "        178     1.0000    1.0000    1.0000         1\n",
    "        179     1.0000    1.0000    1.0000         1\n",
    "        182     1.0000    1.0000    1.0000         4\n",
    "        183     1.0000    1.0000    1.0000         4\n",
    "\n",
    "avg / total     0.9933    0.9932    0.9932    951993\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### CRF\n",
    "\n",
    "```text\n",
    "               precision    recall  f1-score   support\n",
    "\n",
    "         case     0.9584    0.9687    0.9635     11014\n",
    "          obl     0.8045    0.8274    0.8158      5810\n",
    "         flat     0.9469    0.9551    0.9510     10648\n",
    "           cc     0.9538    0.9652    0.9595      3336\n",
    "         conj     0.8684    0.8482    0.8582      4560\n",
    "        punct     0.9848    0.9963    0.9905     17017\n",
    "   nsubj:pass     0.8336    0.7640    0.7973      2059\n",
    "         root     0.7960    0.8453    0.8199      5037\n",
    "       nummod     0.9334    0.9359    0.9347      4088\n",
    "         mark     0.8739    0.8865    0.8802      1392\n",
    "        advcl     0.7649    0.6508    0.7033      1200\n",
    "       advmod     0.8932    0.8924    0.8928      4769\n",
    "         nmod     0.7762    0.7355    0.7553      4215\n",
    "        nsubj     0.8600    0.8835    0.8716      6388\n",
    "          det     0.9020    0.8868    0.8943      4142\n",
    "     compound     0.8776    0.8974    0.8874      6869\n",
    "         amod     0.8677    0.8530    0.8602      4128\n",
    "          obj     0.8749    0.8765    0.8757      5256\n",
    "          acl     0.8375    0.8094    0.8232      3075\n",
    "        xcomp     0.8082    0.8070    0.8076      1264\n",
    "    parataxis     0.7636    0.6208    0.6848       385\n",
    "        appos     0.8221    0.8177    0.8199      2425\n",
    "          cop     0.9350    0.9498    0.9423      1015\n",
    "        fixed     0.8569    0.8056    0.8305       602\n",
    "        ccomp     0.7516    0.5576    0.6402       434\n",
    "compound:plur     0.9154    0.9498    0.9323       638\n",
    "          dep     0.7820    0.5275    0.6300       510\n",
    "        csubj     0.8750    0.8400    0.8571        25\n",
    "         iobj     0.9375    0.6818    0.7895        22\n",
    "   csubj:pass     1.0000    0.8000    0.8889         5\n",
    "          aux     0.5000    0.2500    0.3333         4\n",
    "\n",
    "  avg / total     0.8953    0.8961    0.8953    112332\n",
    "  \n",
    "             precision    recall  f1-score   support\n",
    "\n",
    "          5     0.5452    0.5875    0.5656      5964\n",
    "          2     0.6193    0.7164    0.6643      4365\n",
    "          1     0.8839    0.9031    0.8934      4942\n",
    "          7     0.5181    0.5460    0.5317      5505\n",
    "          9     0.5569    0.5504    0.5536      4804\n",
    "         12     0.5421    0.5309    0.5364      3760\n",
    "         15     0.5556    0.5105    0.5321      3181\n",
    "          4     0.5195    0.6219    0.5661      6241\n",
    "          6     0.5346    0.5571    0.5456      5942\n",
    "         11     0.5350    0.5581    0.5463      4150\n",
    "         14     0.5425    0.5109    0.5262      3251\n",
    "          8     0.5463    0.5414    0.5438      5395\n",
    "         10     0.5705    0.5252    0.5469      4682\n",
    "         13     0.5506    0.5199    0.5348      3537\n",
    "          3     0.5871    0.6077    0.5972      5068\n",
    "         18     0.5613    0.5232    0.5415      2504\n",
    "         20     0.5772    0.5315    0.5534      2109\n",
    "         23     0.6065    0.5814    0.5937      1689\n",
    "         26     0.5820    0.5861    0.5841      1138\n",
    "         29     0.6089    0.5874    0.5980      1047\n",
    "         32     0.6459    0.6241    0.6348       798\n",
    "         35     0.6659    0.5931    0.6274       521\n",
    "         36     0.6312    0.6406    0.6359       537\n",
    "         40     0.6039    0.6620    0.6316       426\n",
    "         17     0.5513    0.5303    0.5406      2674\n",
    "         22     0.5889    0.5238    0.5545      1827\n",
    "         25     0.5898    0.5967    0.5932      1381\n",
    "         27     0.5802    0.5588    0.5693      1088\n",
    "         28     0.6101    0.6082    0.6092       970\n",
    "         34     0.6011    0.6029    0.6020       695\n",
    "         39     0.6711    0.5884    0.6270       430\n",
    "         37     0.6675    0.5876    0.6250       468\n",
    "         42     0.6975    0.6323    0.6633       310\n",
    "         43     0.6504    0.6584    0.6544       243\n",
    "         44     0.7205    0.6221    0.6677       344\n",
    "         47     0.6667    0.7077    0.6866       260\n",
    "         49     0.6903    0.7290    0.7091       214\n",
    "         51     0.6829    0.7368    0.7089       190\n",
    "         53     0.7483    0.6730    0.7086       159\n",
    "         55     0.7143    0.6936    0.7038       173\n",
    "         57     0.7093    0.6224    0.6630        98\n",
    "         59     0.7652    0.6779    0.7189       149\n",
    "         60     0.7253    0.7174    0.7213        92\n",
    "         61     0.7658    0.7516    0.7586       161\n",
    "         62     0.6500    0.5571    0.6000        70\n",
    "         63     0.7257    0.7736    0.7489       106\n",
    "         64     0.8730    0.7971    0.8333        69\n",
    "         65     0.8533    0.6667    0.7485        96\n",
    "         66     0.7097    0.8354    0.7674        79\n",
    "         67     0.5965    0.6415    0.6182        53\n",
    "         72     0.9362    0.6769    0.7857        65\n",
    "         70     0.9024    0.5968    0.7184        62\n",
    "         75     0.9348    0.7414    0.8269        58\n",
    "         77     0.7838    0.8286    0.8056        35\n",
    "         78     0.8750    0.8750    0.8750        16\n",
    "         80     0.7200    0.8000    0.7579        45\n",
    "         82     0.7027    0.7222    0.7123        36\n",
    "         83     0.6923    0.7200    0.7059        25\n",
    "         84     0.7407    0.5000    0.5970        40\n",
    "         85     0.6923    0.8571    0.7660        21\n",
    "         86     0.9091    0.6061    0.7273        33\n",
    "         87     0.5833    0.5000    0.5385        28\n",
    "         88     0.8333    0.4412    0.5769        34\n",
    "         89     0.7619    0.9412    0.8421        17\n",
    "         90     0.9143    0.7805    0.8421        41\n",
    "         91     0.6923    0.8182    0.7500        22\n",
    "         92     1.0000    0.8519    0.9200        27\n",
    "         93     1.0000    0.7273    0.8421        22\n",
    "         94     0.9333    0.6364    0.7568        22\n",
    "         95     1.0000    0.6250    0.7692        16\n",
    "         96     0.8000    0.6857    0.7385        35\n",
    "         97     0.9500    0.8261    0.8837        23\n",
    "        100     1.0000    0.6667    0.8000         6\n",
    "        103     1.0000    0.7857    0.8800        14\n",
    "        104     1.0000    0.6000    0.7500        15\n",
    "        101     0.9574    0.8491    0.9000        53\n",
    "        107     0.8846    0.8214    0.8519        28\n",
    "        112     0.8000    0.8000    0.8000        10\n",
    "        115     1.0000    0.5556    0.7143         9\n",
    "        120     0.6667    0.6667    0.6667         6\n",
    "        122     0.5556    0.7143    0.6250         7\n",
    "        124     1.0000    0.2857    0.4444         7\n",
    "        125     0.2857    0.4000    0.3333         5\n",
    "        126     0.5455    0.3529    0.4286        17\n",
    "        127     0.6667    1.0000    0.8000         4\n",
    "        128     1.0000    0.3500    0.5185        20\n",
    "        129     0.7000    0.7778    0.7368         9\n",
    "        130     0.8667    0.9286    0.8966        14\n",
    "        132     1.0000    0.7143    0.8333         7\n",
    "        133     0.5714    1.0000    0.7273         4\n",
    "        134     1.0000    1.0000    1.0000         2\n",
    "        138     0.9091    0.7692    0.8333        13\n",
    "        147     1.0000    0.5789    0.7333        19\n",
    "        149     0.6667    1.0000    0.8000         2\n",
    "        150     0.6667    1.0000    0.8000         2\n",
    "         21     0.5698    0.5552    0.5624      1940\n",
    "         24     0.5665    0.5503    0.5583      1501\n",
    "         30     0.6101    0.5757    0.5924       905\n",
    "         19     0.5310    0.5163    0.5236      2142\n",
    "         16     0.5312    0.5369    0.5340      2917\n",
    "         38     0.6839    0.6005    0.6395       418\n",
    "         33     0.6255    0.5951    0.6100       741\n",
    "         41     0.6913    0.6341    0.6614       399\n",
    "          0     0.4083    0.0926    0.1510       529\n",
    "         31     0.5924    0.5625    0.5771       752\n",
    "         48     0.6432    0.6010    0.6214       198\n",
    "         50     0.7320    0.6222    0.6727       180\n",
    "         52     0.6685    0.6538    0.6611       182\n",
    "         54     0.7024    0.6705    0.6860       176\n",
    "         68     0.7791    0.6262    0.6943       107\n",
    "         79     0.9020    0.8214    0.8598        56\n",
    "         46     0.8037    0.6187    0.6992       278\n",
    "         56     0.7721    0.7095    0.7394       148\n",
    "         98     0.8000    0.5926    0.6809        27\n",
    "         45     0.6513    0.6804    0.6655       291\n",
    "         73     0.8261    0.7451    0.7835        51\n",
    "        105     0.8571    0.7500    0.8000         8\n",
    "        108     0.9091    0.8333    0.8696        12\n",
    "        110     0.8462    0.7857    0.8148        14\n",
    "        114     0.7778    0.4375    0.5600        16\n",
    "        123     0.7500    0.5000    0.6000         6\n",
    "        135     1.0000    0.5625    0.7200        16\n",
    "        139     0.0000    0.0000    0.0000         1\n",
    "        142     1.0000    0.7500    0.8571         4\n",
    "        146     1.0000    1.0000    1.0000         3\n",
    "        151     1.0000    0.5000    0.6667         2\n",
    "         76     0.8400    0.7000    0.7636        30\n",
    "         58     0.6838    0.7207    0.7018       111\n",
    "         69     0.6824    0.7838    0.7296        74\n",
    "         74     0.8605    0.8043    0.8315        46\n",
    "         71     0.8077    0.7778    0.7925        81\n",
    "        109     0.8889    0.7273    0.8000        11\n",
    "         99     0.8889    0.6667    0.7619        12\n",
    "        117     1.0000    0.1429    0.2500         7\n",
    "        116     0.6000    0.6667    0.6316         9\n",
    "        113     0.5833    0.2917    0.3889        24\n",
    "        121     0.7500    0.5000    0.6000         6\n",
    "        131     0.8333    1.0000    0.9091         5\n",
    "        137     1.0000    0.7500    0.8571         4\n",
    "         81     0.9375    0.6522    0.7692        46\n",
    "        118     0.5000    0.5000    0.5000         6\n",
    "        111     0.6000    0.6000    0.6000         5\n",
    "        102     1.0000    0.7143    0.8333         7\n",
    "        106     1.0000    0.7727    0.8718        22\n",
    "        136     0.7778    0.2800    0.4118        25\n",
    "        140     1.0000    0.5000    0.6667         2\n",
    "        141     0.0000    0.0000    0.0000         5\n",
    "        144     1.0000    0.3000    0.4615        10\n",
    "        152     0.0000    0.0000    0.0000         1\n",
    "        153     0.0000    0.0000    0.0000         1\n",
    "        158     1.0000    0.6250    0.7692         8\n",
    "        156     0.9091    0.4762    0.6250        21\n",
    "        160     1.0000    1.0000    1.0000         2\n",
    "        164     1.0000    1.0000    1.0000         2\n",
    "        143     0.0000    0.0000    0.0000         0\n",
    "        155     0.0000    0.0000    0.0000         0\n",
    "        157     0.5000    0.2500    0.3333         4\n",
    "        161     0.0000    0.0000    0.0000         0\n",
    "        162     1.0000    0.2500    0.4000        12\n",
    "        166     0.0000    0.0000    0.0000         0\n",
    "        175     0.0000    0.0000    0.0000         0\n",
    "        173     0.0000    0.0000    0.0000         0\n",
    "        176     1.0000    1.0000    1.0000        16\n",
    "        177     1.0000    1.0000    1.0000         8\n",
    "        178     1.0000    1.0000    1.0000         4\n",
    "        181     0.0000    0.0000    0.0000         0\n",
    "        182     1.0000    1.0000    1.0000        16\n",
    "        119     1.0000    0.7143    0.8333        21\n",
    "        148     0.0000    0.0000    0.0000         0\n",
    "        154     1.0000    0.7500    0.8571         8\n",
    "        159     1.0000    0.2500    0.4000         4\n",
    "        163     0.0000    0.0000    0.0000         8\n",
    "        167     1.0000    1.0000    1.0000        16\n",
    "        174     1.0000    1.0000    1.0000         8\n",
    "        179     1.0000    1.0000    1.0000         4\n",
    "        183     1.0000    1.0000    1.0000        16\n",
    "        145     0.0000    0.0000    0.0000         0\n",
    "\n",
    "avg / total     0.5859    0.5847    0.5836    109699\n",
    "```"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}