paper_experiments/3b - Machine Learning Only Fund Flow Features.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Machine Learning (Only Fund Flow Features)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modules and functions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import all the modules:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "from xgboost import XGBClassifier\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "plt.style.use(\"ggplot\")\n",
    "\n",
    "from notebooks import *"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_file_path = \"dataset-filtered.csv\"\n",
    "df = pd.read_csv(df_file_path, low_memory=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dataset has 158863 rows and 434 columns\n"
     ]
    }
   ],
   "source": [
    "print_dimensions(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracted values:\n",
      "addresses (158863,)\n",
      "features (158863, 73)\n",
      "labels_binary (158863,)\n",
      "labels_multi (158863,)\n"
     ]
    }
   ],
   "source": [
    "addresses, X, y_binary, y_multi, scaler, feature_names = extract_experiment_data(df, [\"fund flow\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "symbol_22\n",
      "symbol_33\n",
      "symbol_39\n",
      "symbol_41\n",
      "symbol_43\n",
      "symbol_49\n",
      "symbol_51\n",
      "symbol_53\n",
      "symbol_54\n",
      "symbol_55\n",
      "symbol_57\n",
      "symbol_59\n",
      "symbol_61\n",
      "symbol_62\n",
      "symbol_63\n",
      "symbol_65\n",
      "symbol_66\n",
      "symbol_67\n",
      "symbol_69\n",
      "symbol_71\n",
      "symbol_73\n",
      "symbol_75\n",
      "symbol_76\n",
      "symbol_77\n",
      "symbol_78\n",
      "symbol_79\n",
      "symbol_81\n",
      "symbol_82\n",
      "symbol_83\n",
      "symbol_85\n",
      "symbol_86\n",
      "symbol_87\n",
      "symbol_93\n",
      "symbol_95\n",
      "symbol_103\n",
      "symbol_105\n",
      "symbol_113\n",
      "symbol_121\n",
      "symbol_122\n",
      "symbol_123\n",
      "symbol_127\n",
      "symbol_129\n",
      "symbol_133\n",
      "symbol_135\n",
      "symbol_137\n",
      "symbol_171\n",
      "symbol_173\n",
      "symbol_177\n",
      "symbol_179\n",
      "symbol_181\n",
      "symbol_183\n",
      "symbol_185\n",
      "symbol_187\n",
      "symbol_189\n",
      "symbol_191\n",
      "symbol_193\n",
      "symbol_194\n",
      "symbol_195\n",
      "symbol_197\n",
      "symbol_198\n",
      "symbol_199\n",
      "symbol_201\n",
      "symbol_202\n",
      "symbol_203\n",
      "symbol_204\n",
      "symbol_205\n",
      "symbol_207\n",
      "symbol_208\n",
      "symbol_209\n",
      "symbol_211\n",
      "symbol_213\n",
      "symbol_214\n",
      "symbol_215\n"
     ]
    }
   ],
   "source": [
    "for feature_name in feature_names:\n",
    "    print(feature_name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_scale_pos_weight = compute_scale_pos_weight(y_binary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_xgb_model():\n",
    "    return XGBClassifier(n_jobs=10,\n",
    "                         scale_pos_weight=xgb_scale_pos_weight,\n",
    "                         n_estimators=25,\n",
    "                         max_depth=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train ROC AUC 0.954 TN  137821 FP  4895 FN    15 TP   245\n",
      "test  ROC AUC 0.924 TN  15256 FP   596 FN     4 TP    31\n",
      "train score - test score = 0.030\n",
      "\n",
      "train ROC AUC 0.950 TN  137160 FP  5552 FN    16 TP   248\n",
      "test  ROC AUC 0.948 TN  15244 FP   612 FN     2 TP    29\n",
      "train score - test score = 0.002\n",
      "\n",
      "train ROC AUC 0.953 TN  137508 FP  5206 FN    15 TP   247\n",
      "test  ROC AUC 0.936 TN  15268 FP   586 FN     3 TP    30\n",
      "train score - test score = 0.017\n",
      "\n",
      "train ROC AUC 0.954 TN  137137 FP  5576 FN    14 TP   250\n",
      "test  ROC AUC 0.931 TN  15209 FP   646 FN     3 TP    28\n",
      "train score - test score = 0.023\n",
      "\n",
      "train ROC AUC 0.951 TN  138504 FP  4212 FN    18 TP   243\n",
      "test  ROC AUC 0.928 TN  15434 FP   418 FN     4 TP    30\n",
      "train score - test score = 0.023\n",
      "\n",
      "train ROC AUC 0.952 TN  137491 FP  5219 FN    16 TP   251\n",
      "test  ROC AUC 0.947 TN  15297 FP   561 FN     2 TP    26\n",
      "train score - test score = 0.005\n",
      "\n",
      "train ROC AUC 0.949 TN  138545 FP  4170 FN    19 TP   243\n",
      "test  ROC AUC 0.925 TN  15410 FP   443 FN     4 TP    29\n",
      "train score - test score = 0.024\n",
      "\n",
      "train ROC AUC 0.956 TN  137407 FP  5291 FN    14 TP   265\n",
      "test  ROC AUC 0.888 TN  15288 FP   582 FN     3 TP    13\n",
      "train score - test score = 0.068\n",
      "\n",
      "train ROC AUC 0.949 TN  137696 FP  5010 FN    18 TP   253\n",
      "test  ROC AUC 0.982 TN  15296 FP   566 FN     0 TP    24\n",
      "train score - test score = -0.033\n",
      "\n",
      "train ROC AUC 0.950 TN  137464 FP  5248 FN    17 TP   248\n",
      "test  ROC AUC 0.948 TN  15255 FP   601 FN     2 TP    28\n",
      "train score - test score = 0.002\n",
      "\n",
      "train: 0.952 +- 0.002 test: 0.936 +- 0.023\n",
      "\n"
     ]
    }
   ],
   "source": [
    "xgb_models = train_test_folds(X,\n",
    "                              y_binary,\n",
    "                              k_fold(X, n_splits=10),\n",
    "                              create_xgb_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_importance = compute_average_feature_importance(X, xgb_models)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature</th>\n",
       "      <th>Importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>symbol_83</td>\n",
       "      <td>0.797042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>symbol_201</td>\n",
       "      <td>0.036533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>symbol_79</td>\n",
       "      <td>0.031031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>symbol_39</td>\n",
       "      <td>0.027004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>symbol_73</td>\n",
       "      <td>0.023639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>symbol_205</td>\n",
       "      <td>0.015515</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>symbol_33</td>\n",
       "      <td>0.015306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>symbol_213</td>\n",
       "      <td>0.010379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>symbol_77</td>\n",
       "      <td>0.010184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>symbol_173</td>\n",
       "      <td>0.008347</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Feature  Importance\n",
       "0   symbol_83    0.797042\n",
       "1  symbol_201    0.036533\n",
       "2   symbol_79    0.031031\n",
       "3   symbol_39    0.027004\n",
       "4   symbol_73    0.023639\n",
       "5  symbol_205    0.015515\n",
       "6   symbol_33    0.015306\n",
       "7  symbol_213    0.010379\n",
       "8   symbol_77    0.010184\n",
       "9  symbol_173    0.008347"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(create_feature_importance_table(feature_names, feature_importance, size=10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_feature_importance(feature_names, feature_importance)"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}