NEW_ASSET_ARIMA_MODELING.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import random\n",
    "from pyramid.arima import ARIMA, auto_arima\n",
    "from statsmodels.graphics.tsaplots import plot_acf, plot_pacf\n",
    "import warnings\n",
    "warnings.simplefilter(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "dataset = pd.read_csv('C:/Users/Froilan/Desktop/myFiles/JupyterFiles/stock_correlation_prediction/train_dev_test/new_asset_before_arima.csv')\n",
    "dataset = dataset.loc[:,~dataset.columns.str.contains('Unnamed')]\n",
    "\n",
    "model_110 = ARIMA(order=(1,1,0), method='mle', suppress_warnings=True)\n",
    "model_011 = ARIMA(order=(0,1,1), method='mle', suppress_warnings=True)\n",
    "model_111 = ARIMA(order=(1,1,1), method='mle', suppress_warnings=True)\n",
    "model_211 = ARIMA(order=(2,1,1), method='mle', suppress_warnings=True)\n",
    "model_210 = ARIMA(order=(2,1,0), method='mle', suppress_warnings=True)\n",
    "\n",
    "flag = 0\n",
    "c=0\n",
    "residual = []\n",
    "for s in np.array(dataset):\n",
    "    c+=1\n",
    "    try:\n",
    "        model1 = model_110.fit(s)\n",
    "        model = model1\n",
    "\n",
    "        try:\n",
    "            model2 = model_011.fit(s)\n",
    "\n",
    "            if model.aic() <= model2.aic() :\n",
    "                pass\n",
    "            else :\n",
    "                model = model2\n",
    "\n",
    "            try :\n",
    "                model3 = model_111.fit(s)\n",
    "                if model.aic() <= model3.aic() :\n",
    "                    pass\n",
    "                else :\n",
    "                    model = model3\n",
    "            except :\n",
    "                try:\n",
    "                    model4 = model_211.fit(s)\n",
    "\n",
    "                    if model.aic() <= model4.aic() :\n",
    "                        pass\n",
    "                    else:\n",
    "                        model = model4\n",
    "                except:\n",
    "                    try:\n",
    "                        model5 = model_210.fit(s)\n",
    "\n",
    "                        if model.aic() <= model5.aic():\n",
    "                            pass\n",
    "                        else :\n",
    "                            model = model5\n",
    "                    except :\n",
    "                        pass\n",
    "\n",
    "        except:\n",
    "            try:\n",
    "                model3 = model_111.fit(s)\n",
    "\n",
    "                if model.aic() <= model3.aic() :\n",
    "                    pass\n",
    "                else :\n",
    "                    model = model3\n",
    "            except :\n",
    "                try:\n",
    "                    model4 = model_211.fit(s)\n",
    "\n",
    "                    if model.aic() <= model4.aic() :\n",
    "                        pass\n",
    "                    else:\n",
    "                        model = model4\n",
    "                except:\n",
    "                    try:\n",
    "                        model5 = model_210.fit(s)\n",
    "\n",
    "                        if model.aic() <= model5.aic():\n",
    "                            pass\n",
    "                        else :\n",
    "                            model = model5\n",
    "                    except :\n",
    "                        pass\n",
    "\n",
    "    except:\n",
    "        try:\n",
    "            model2 = model_011.fit(s[i])\n",
    "            model = model2\n",
    "\n",
    "            try :\n",
    "                model3 = model_111.fit(s[i])\n",
    "\n",
    "                if model.aic() <= model3.aic():\n",
    "                    pass\n",
    "                else:\n",
    "                    model = model3\n",
    "            except :\n",
    "                try:\n",
    "                    model4 = model_211.fit(s[i])\n",
    "\n",
    "                    if model.aic() <= model4.aic() :\n",
    "                        pass\n",
    "                    else:\n",
    "                        model = model4\n",
    "                except:\n",
    "                    try:\n",
    "                        model5 = model_210.fit(s[i])\n",
    "\n",
    "                        if model.aic() <= model5.aic():\n",
    "                            pass\n",
    "                        else :\n",
    "                            model = model5\n",
    "                    except :\n",
    "                        pass\n",
    "\n",
    "        except :\n",
    "            try:\n",
    "                model3 = model_111.fit(s[i])\n",
    "                model = model3\n",
    "            except :\n",
    "                try:\n",
    "                    model4 = model_211.fit(s[i])\n",
    "\n",
    "                    if model.aic() <= model4.aic() :\n",
    "                        pass\n",
    "                    else:\n",
    "                        model = model4\n",
    "                except:\n",
    "                    try:\n",
    "                        model5 = model_210.fit(s[i])\n",
    "\n",
    "                        if model.aic() <= model5.aic():\n",
    "                            pass\n",
    "                        else :\n",
    "                            model = model5\n",
    "                    except :\n",
    "                        flag = 1\n",
    "                        print(str(c) + \" FATAL ERROR\")\n",
    "                        break\n",
    "\n",
    "                        \n",
    "    predictions = list(model.predict_in_sample())\n",
    "\n",
    "    predictions = [np.mean(predictions)] + predictions\n",
    "\n",
    "    res = pd.Series(np.array(s) - np.array(predictions))\n",
    "    residual.append(np.array(res))\n",
    "\n",
    "    if flag == 1:\n",
    "        break\n",
    "residual = pd.DataFrame(residual)\n",
    "residual.to_csv('C:/Users/Froilan/Desktop/myFiles/JupyterFiles/stock_correlation_prediction/train_dev_test/new_asset_after_arima.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}