unemp_anal_RFR_v02_171021.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# USA Unemployment analysis with ScikitLearn RF Regressor\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<script>\n",
       "code_show=true; \n",
       "function code_toggle() {\n",
       " if (code_show){\n",
       " $('div.input').hide();\n",
       " } else {\n",
       " $('div.input').show();\n",
       " }\n",
       " code_show = !code_show\n",
       "} \n",
       "$( document ).ready(code_toggle);\n",
       "</script>\n",
       "<form action=\"javascript:code_toggle()\"><input type=\"submit\" value=\"Toggle code.\"></form>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "from IPython.display import HTML\n",
    "\n",
    "HTML('''<script>\n",
    "code_show=true; \n",
    "function code_toggle() {\n",
    " if (code_show){\n",
    " $('div.input').hide();\n",
    " } else {\n",
    " $('div.input').show();\n",
    " }\n",
    " code_show = !code_show\n",
    "} \n",
    "$( document ).ready(code_toggle);\n",
    "</script>\n",
    "<form action=\"javascript:code_toggle()\"><input type=\"submit\" value=\"Toggle code.\"></form>''')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 0. Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from IPython.display import display\n",
    "from IPython.core.pylabtools import figsize, getfigs\n",
    "%matplotlib inline\n",
    "\n",
    "import sklearn\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "from numpy import corrcoef\n",
    "from time import clock\n",
    "\n",
    "from sklearn import tree\n",
    "from sklearn.tree import export_graphviz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "\n",
    "\n",
    "## 1. Data input\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         pop  unempRateWords\n",
      "fips                        \n",
      "1003  200111          5.9250\n",
      "1005   26887         11.4750\n",
      "1007   22506          7.1000\n",
      "1009   57719          6.0625\n",
      "1011   10764         11.7500 \n",
      "\n",
      " (2956, 2)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  Unemployment rates  //words dataset\n",
    "\n",
    "unemp = pd.read_csv(\"unemp_small.csv\", sep=\"\\t\", header = None)\n",
    "unemp.columns = [\"fips\", \"pop\", \"unempRateWords\"]\n",
    "\n",
    "unempPivot = unemp.pivot_table(index=\"fips\",values=[\"unempRateWords\",\"pop\"]).copy()\n",
    "unempPivot.drop_duplicates(subset=\"pop\", keep=False, inplace=True)\n",
    "\n",
    "print(unempPivot.head(), '\\n\\n', unempPivot.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      employed  unempRateAct\n",
      "fips                        \n",
      "1001     24111           5.9\n",
      "1003     81131           6.1\n",
      "1005      7917          10.6\n",
      "1007      7946           7.2\n",
      "1009     23036           6.1 \n",
      "\n",
      " (3219, 2)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Program Files\\Python\\lib\\site-packages\\ipykernel_launcher.py:5: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.\n",
      "  \"\"\"\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  Unemployment rates  //acticity dataset\n",
    "\n",
    "unemp2 = pd.read_table(\"laucnty14.txt\",sep=r'\\s{2,}',header=None,usecols=[1,2,5,6,7,8,9],\n",
    "         converters={1: np.str, 2: np.str, 8: np.float64})\n",
    "#print(unemp2.head(),'\\n\\n')\n",
    "\n",
    "unemp2.columns = [\"fip1\", \"fip2\", \"laborForce\", \"employed\", \"unemployed\", \"unempRateAct\"]\n",
    "unemp2[\"fips\"] = unemp2[\"fip1\"] + unemp2[\"fip2\"]\n",
    "unemp2[\"fips\"] = unemp2[\"fips\"].astype(int)\n",
    "unemp2[\"laborForce\"] = unemp2[\"laborForce\"].apply(lambda s: s.replace(',','')).astype(int)\n",
    "unemp2[\"unemployed\"] = unemp2[\"unemployed\"].apply(lambda s: s.replace(',','')).astype(int)\n",
    "unemp2[\"employed\"] = unemp2[\"employed\"].apply(lambda s: s.replace(',','')).astype(int)\n",
    "\n",
    "unemp2Pivot = unemp2.pivot_table(index=\"fips\", values=[\"unempRateAct\",\"employed\"]).copy()\n",
    "\n",
    "print(unemp2Pivot.head(), '\\n\\n', unemp2Pivot.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "        empRate  unempRateAct  unempRateWords\n",
      "fips                                         \n",
      "1003  40.542999           6.1          5.9250\n",
      "1005  29.445457          10.6         11.4750\n",
      "1007  35.306141           7.2          7.1000\n",
      "1009  39.910601           6.1          6.0625\n",
      "1011  40.096618           8.8         11.7500 \n",
      "\n",
      " (2954, 3)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  Unemployment Data  //merge\n",
    "\n",
    "# ! Unemployment rate based on labor force, employment rate based on population !\n",
    "# ! activity dataset and words dataset has different unemployment rates !\n",
    "\n",
    "unempData = pd.merge(unempPivot,unemp2Pivot, how='inner', left_index=True, right_index=True).copy()\n",
    "#print(unempData.head(),'\\n\\n')\n",
    "\n",
    "unempData[\"empRate\"] = 100 * unempData[\"employed\"] / unempData[\"pop\"]\n",
    "unempData.drop([\"pop\",\"employed\"], axis=1, inplace=True)\n",
    "unempData.sort_index(level=0, axis=0, ascending=True, inplace=True, sort_remaining=False)\n",
    "unempData.sort_index(level=0, axis=1, ascending=True, inplace=True, sort_remaining=False)\n",
    "\n",
    "print(unempData.head(), '\\n\\n', unempData.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    fips     word  cnt   totcnt      freq\n",
      "0  48473    aaron    2    95626  0.000021\n",
      "1  39095     abby   17   816992  0.000021\n",
      "2  34033  ability    5   113270  0.000044\n",
      "3   1127    aaron    5    90625  0.000055\n",
      "4   6071     abby   30  3152613  0.000010 \n",
      "\n",
      " (6557665, 5)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  Word frequency Data (BIG FILE !!!)\n",
    "\n",
    "words = pd.read_csv(\"unemp_training_set.csv\", header = None)\n",
    "words.columns = [\"fips\", \"word\", \"cnt\", \"totcnt\"]\n",
    "words[\"word\"] = words[\"word\"].apply(lambda s: s.lower())\n",
    "words[\"freq\"] = words.apply(lambda row: row[\"cnt\"]/row[\"totcnt\"], axis = 1)\n",
    "\n",
    "print(words.head(), '\\n\\n', words.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "        empRate  unempRateWords     aaron      abby   ability      able  \\\n",
      "fips                                                                      \n",
      "1003  40.542999          5.9250  0.000042  0.000021  0.000017  0.000431   \n",
      "1009  39.910601          6.0625  0.000026  0.000052  0.000052  0.000442   \n",
      "1013  41.791486          8.7875  0.000000  0.000034  0.000034  0.000309   \n",
      "1015  37.520273          7.7000  0.000014  0.000024  0.000005  0.000360   \n",
      "1017  41.921000          7.1125  0.000000  0.000015  0.000015  0.000362   \n",
      "\n",
      "         about     above  absolute  absolutely    ...         youu     yummy  \\\n",
      "fips                                              ...                          \n",
      "1003  0.005418  0.000088  0.000029    0.000222    ...     0.000050  0.000038   \n",
      "1009  0.006998  0.000000  0.000026    0.000364    ...     0.000104  0.000000   \n",
      "1013  0.006377  0.000240  0.000000    0.000069    ...     0.000000  0.000000   \n",
      "1015  0.005733  0.000072  0.000077    0.000163    ...     0.000091  0.000019   \n",
      "1017  0.006019  0.000106  0.000030    0.000106    ...     0.000196  0.000015   \n",
      "\n",
      "          yung      zach      zack      zayn      zero    zombie   zombies  \\\n",
      "fips                                                                         \n",
      "1003  0.000100  0.000100  0.000033  0.000000  0.000071  0.000021  0.000004   \n",
      "1009  0.000000  0.000000  0.000000  0.000026  0.000078  0.000000  0.000026   \n",
      "1013  0.000034  0.000069  0.000000  0.000000  0.000034  0.000000  0.000034   \n",
      "1015  0.000048  0.000062  0.000010  0.000000  0.000048  0.000005  0.000010   \n",
      "1017  0.000015  0.000015  0.000045  0.000000  0.000045  0.000015  0.000000   \n",
      "\n",
      "          zone  \n",
      "fips            \n",
      "1003  0.000084  \n",
      "1009  0.000052  \n",
      "1013  0.000034  \n",
      "1015  0.000086  \n",
      "1017  0.000151  \n",
      "\n",
      "[5 rows x 5987 columns] \n",
      "\n",
      " (1301, 5987)\n"
     ]
    }
   ],
   "source": [
    "wordsPivot = words.pivot_table(columns=\"word\",index=\"fips\",values=\"freq\").copy()\n",
    "wordsPivot.fillna(0, inplace=True)\n",
    "wordsPivot.sort_index(level=0, axis=0, ascending=True, inplace=True, sort_remaining=False)\n",
    "wordsPivot.sort_index(level=0, axis=1, ascending=True, inplace=True, sort_remaining=False)\n",
    "#print(wordsPivot.head(),'\\n\\n')\n",
    "\n",
    "wordsData = pd.merge(unempData ,wordsPivot, how='inner', left_index=True, right_index=True).copy()\n",
    "wordsData.drop(\"unempRateAct\", axis=1, inplace=True)\n",
    "wordsData.sort_index(level=0, axis=0, ascending=True, inplace=True, sort_remaining=False)\n",
    "\n",
    "print(wordsData.head(), '\\n\\n', wordsData.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "        empRate  unempRateAct         0         1         2         3  \\\n",
      "fips                                                                    \n",
      "1003  40.542999           6.1  0.068764  0.048920  0.027950  0.015420   \n",
      "1007  35.306141           7.2  0.049163  0.040795  0.030858  0.031381   \n",
      "1009  39.910601           6.1  0.057902  0.038271  0.022435  0.008743   \n",
      "1013  41.791486           8.6  0.055371  0.036446  0.022954  0.012616   \n",
      "1015  37.520273           8.0  0.074314  0.064947  0.043261  0.027678   \n",
      "\n",
      "             4         5         6         7      ...             15  \\\n",
      "fips                                              ...                  \n",
      "1003  0.008567  0.005472  0.011456  0.020228      ...       0.038333   \n",
      "1007  0.020921  0.007845  0.019874  0.022490      ...       0.042887   \n",
      "1009  0.005279  0.005444  0.016331  0.019630      ...       0.041405   \n",
      "1013  0.005782  0.006133  0.019450  0.023655      ...       0.043806   \n",
      "1015  0.013726  0.008691  0.010576  0.015330      ...       0.035272   \n",
      "\n",
      "            16        17        18        19        20        21        22  \\\n",
      "fips                                                                         \n",
      "1003  0.040020  0.043805  0.049942  0.057282  0.062524  0.070093  0.078276   \n",
      "1007  0.052824  0.035565  0.063285  0.078975  0.075314  0.075837  0.077406   \n",
      "1009  0.046684  0.048499  0.061531  0.064500  0.078522  0.092379  0.102936   \n",
      "1013  0.050464  0.040652  0.053969  0.064132  0.077274  0.093569  0.080252   \n",
      "1015  0.034400  0.042051  0.048352  0.052261  0.059828  0.063428  0.077605   \n",
      "\n",
      "            23  messageCount  \n",
      "fips                          \n",
      "1003  0.079095       39105.0  \n",
      "1007  0.064331        1912.0  \n",
      "1009  0.075388        6062.0  \n",
      "1013  0.065183        5707.0  \n",
      "1015  0.081543       35552.0  \n",
      "\n",
      "[5 rows x 27 columns] \n",
      "\n",
      " (1823, 27)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  Activity data\n",
    "\n",
    "activ = pd.read_csv(\"raw_data.csv\", sep=\",\", header = None)\n",
    "activ.columns = [\"fips\", \"hour\", \"activity\"]\n",
    "#print(activ.head(), \"\\n\\n\")\n",
    "\n",
    "activPivot = activ.pivot_table(index=\"fips\", columns=\"hour\", values=\"activity\").copy()\n",
    "activPivot.fillna(0, inplace=True)\n",
    "activPivot.sort_index(level=0, axis=1, ascending=True, inplace=True, sort_remaining=False)\n",
    "#print(activPivot.head(), \"\\n\\n\")\n",
    "\n",
    "activPivot[\"messageCount\"] = activPivot.sum(numeric_only=True, axis=1)\n",
    "activPivot = activPivot[activPivot[\"messageCount\"] > 1800]\n",
    "activPivot.iloc[:,:-1] = activPivot.iloc[:,:-1].div(activPivot[\"messageCount\"], axis=0)\n",
    "#print(activPivot.head(), \"\\n\\n\")\n",
    "\n",
    "activData = pd.merge(unempData ,activPivot, how='inner', left_index=True, right_index=True).copy()\n",
    "activData.drop(\"unempRateWords\", axis=1, inplace=True)\n",
    "activData.sort_index(level=0, axis=0, ascending=True, inplace=True, sort_remaining=False)\n",
    "\n",
    "print(activData.head(), '\\n\\n', activData.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Data for ML"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "labelWordsUnemp:  [ 5.925   6.0625  8.7875  7.7     7.1125] (1301,) \n",
      "\n",
      "labelWordsEmp:  [ 40.54299864  39.91060136  41.79148601  37.5202733   41.92100012] (1301,) \n",
      "\n",
      "featuresWords_list:  ['aaron' 'abby' 'ability' ..., 'zombie' 'zombies' 'zone'] (5985,) \n",
      "\n",
      "featuresWords: \n",
      " [[  4.18366280e-05   2.09183140e-05   1.67346512e-05]\n",
      " [  2.60132147e-05   5.20264294e-05   5.20264294e-05]\n",
      " [  0.00000000e+00   3.42829716e-05   3.42829716e-05]\n",
      " [  1.44050014e-05   2.40083357e-05   4.80166714e-06]\n",
      " [  0.00000000e+00   1.50847764e-05   1.50847764e-05]] (1301, 5985) \n",
      "\n",
      "fipsWords_list:  [1003 1009 1013 1015 1017] (1301,) \n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  WordsData\n",
    "\n",
    "labelWordsUnemp = (wordsData.unempRateWords.as_matrix()).copy()\n",
    "print (\"labelWordsUnemp: \", labelWordsUnemp[:5], labelWordsUnemp.shape,\"\\n\")\n",
    "\n",
    "labelWordsEmp = (wordsData.empRate.as_matrix()).copy()\n",
    "print (\"labelWordsEmp: \", labelWordsEmp[:5], labelWordsEmp.shape,\"\\n\")\n",
    "\n",
    "featuresWords_list = wordsData.columns.values[2:].copy()\n",
    "print (\"featuresWords_list: \", featuresWords_list, featuresWords_list.shape,\"\\n\")\n",
    "\n",
    "featuresWords = (wordsData.iloc[:,2:].as_matrix()).copy()\n",
    "print (\"featuresWords: \\n\", featuresWords[:5,:3], featuresWords.shape,\"\\n\")\n",
    "\n",
    "fipsWords_list = (wordsData.index.values[:]).copy()\n",
    "print (\"fipsWords_list: \", fipsWords_list[:5], fipsWords_list.shape,\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "labelActivUnemp:  [ 6.1  7.2  6.1  8.6  8. ] (1823,) \n",
      "\n",
      "labelActivEmp:  [ 40.54299864  35.30614058  39.91060136  41.79148601  37.5202733 ] (1823,) \n",
      "\n",
      "featuresActiv_list:  [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23] (24,) \n",
      "\n",
      "featuresActiv: \n",
      " [[ 0.06876359  0.04891958  0.02795039]\n",
      " [ 0.04916318  0.04079498  0.03085774]\n",
      " [ 0.05790168  0.0382712   0.02243484]\n",
      " [ 0.0553706   0.03644647  0.02295427]\n",
      " [ 0.07431368  0.06494712  0.04326058]] (1823, 24) \n",
      "\n",
      "fipsActiv_list:  [1003 1007 1009 1013 1015] (1823,) \n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "##  ActivData\n",
    "\n",
    "labelActivUnemp = (activData.unempRateAct.as_matrix()).copy()\n",
    "print (\"labelActivUnemp: \", labelActivUnemp[:5], labelActivUnemp.shape,\"\\n\")\n",
    "\n",
    "labelActivEmp = (activData.empRate.as_matrix()).copy()\n",
    "print (\"labelActivEmp: \", labelActivEmp[:5], labelActivEmp.shape,\"\\n\")\n",
    "\n",
    "featuresActiv_list = (activData.columns.values[2:-1]).copy()\n",
    "print (\"featuresActiv_list: \", featuresActiv_list, featuresActiv_list.shape,\"\\n\")\n",
    "\n",
    "featuresActiv = (activData.iloc[:,2:-1].as_matrix()).copy()\n",
    "print (\"featuresActiv: \\n\", featuresActiv[:5,:3], featuresActiv.shape,\"\\n\")\n",
    "\n",
    "fipsActiv_list = (activData.index.values[:]).copy()\n",
    "print (\"fipsActiv_list: \", fipsActiv_list[:5], fipsActiv_list.shape,\"\\n\")\n",
    "\n",
    "#weight = (activData.iloc[:,-1].as_matrix()).copy()\n",
    "#print (\"weight: \", weight[:5], weight.shape,\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "##  All output data:\n",
    "\n",
    "# labelWordsUnemp\n",
    "# labelWordsEmp\n",
    "# featuresWords_list\n",
    "# featuresWords\n",
    "# fipsWords_list\n",
    "\n",
    "# labelActivUnemp\n",
    "# labelActivEmp\n",
    "# featuresActiv_list\n",
    "# featuresActiv\n",
    "# fipsActiv_list"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Machine Learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RF Regressor auto parameter tester"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "## Automatizalt parameter teszteles\n",
    "\n",
    "def autoParamSearch(features,labels,nEstimators,maxDepth,maxFeatures,file):\n",
    "    \n",
    "    X = features.copy()\n",
    "    y = labels.copy()\n",
    "    \n",
    "    for nE in nEstimators:\n",
    "        for mD in maxDepth:\n",
    "            for mF in maxFeatures:\n",
    "                \n",
    "                regr_rf = RandomForestRegressor(n_estimators=nE,max_depth=mD,max_features=mF,n_jobs=3)\n",
    "                corr = []\n",
    "                time = []\n",
    "                \n",
    "                for i in range(0,10):    \n",
    "                    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5)\n",
    "                    t0 = clock()\n",
    "                    regr_rf.fit(X_train, y_train)\n",
    "                    t1 = clock()\n",
    "                    y_pred = regr_rf.predict(X_test)\n",
    "                    corr.append(corrcoef(y_test,y_pred)[0,1])\n",
    "                    time.append(t1-t0)\n",
    "                    \n",
    "                corr = np.array(corr)\n",
    "                err = (corr.max() - corr.min()) / 2\n",
    "                time = np.array(time)\n",
    "                file.write(\"%f %f %f %d %d %d\\n\" % (corr.mean(), err, time.mean(), nE, mD, mF))\n",
    "                \n",
    "    return 0\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "'''\n",
    "%%time\n",
    "\n",
    "## Parameterek tesztelese\n",
    "\n",
    "f = open(\"test.txt\",\"w\")\n",
    "\n",
    "nEstimators = [100]      #[100,1000]\n",
    "maxDepth = [12]         #[4,5,6,8,10,14,18]\n",
    "maxFeatures = [100]      #[80,1556,3032,4508,5985] # [6,8,10,12,14,16,18,20,22,24]\n",
    "\n",
    "returnVal = autoParamSearch(featuresWords,labelWordsEmp,nEstimators,maxDepth,maxFeatures,f)\n",
    "print(\"returnVal: \", returnVal)\n",
    "\n",
    "f.close()\n",
    "''';"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RF Regressor for manual test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "X = featuresActiv.copy()\n",
    "y = labelActivEmp.copy()\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correaltion:  0.440761161715 \n",
      "\n",
      "Wall time: 4.14 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "regr_rf = RandomForestRegressor(n_estimators=1000, max_depth=12, max_features=5, n_jobs=3)\n",
    "regr_rf.fit(X_train, y_train)\n",
    "y_pred = regr_rf.predict(X_test)\n",
    "\n",
    "print(\"Correaltion: \", corrcoef(y_test,y_pred)[0,1], \"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Tree graph export to a .dot file (manual convert to png from termianl: \"dot -Tpng tree.dot -o tree.png\"\n",
    "#                                   require dot.exe (package dont inculde dot.exe) )\n",
    "'''\n",
    "tree.export_graphviz(regr_rf.estimators_[0], feature_names=featuresActiv_list, filled=True, rounded=True, out_file='tree.dot')\n",
    "''';"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predicted values plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAusAAALlCAYAAAB986hNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xa80d2e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Init\n",
    "\n",
    "project_name = \"test\"\n",
    "label = \"(Un)employment rate\"\n",
    "realVal = y_test\n",
    "predVal = y_pred\n",
    "\n",
    "## Plot (real value, predicted value)\n",
    "\n",
    "# fit_fn is a function which takes in x and returns an estimate for y\n",
    "fit = np.polyfit(realVal,predVal,1)\n",
    "fit_fn = np.poly1d(fit)\n",
    "x = np.linspace(0,100,10000) \n",
    "\n",
    "figsize(12,12)\n",
    "plt.plot(realVal,predVal,color='b',marker='.',ls='',ms=3,label=label)\n",
    "plt.plot(x,fit_fn(x),color='g',ls='--',lw=2,label='linear fit')\n",
    "plt.plot(x,x,color='k',ls='--',lw=2,label='x = y line')\n",
    "\n",
    "plt.title('%s prediction %s' % (label,project_name),fontsize=25)\n",
    "plt.xlabel('Real value',fontsize=25) \n",
    "plt.ylabel('Predicted value',fontsize=25)\n",
    "plt.legend(loc='lower right',fontsize=25)\n",
    "plt.xticks(fontsize=15)\n",
    "plt.yticks(fontsize=15)\n",
    "plt.grid(True)\n",
    "\n",
    "plt.xlim(0,70)\n",
    "plt.ylim(0,70);\n",
    "\n",
    "# plt.text(2, 11, 'Corr = 0.44 +- 0.06',color='b',fontsize=15)\n",
    "# plt.savefig(project_name+\"UnempPred.png\",dpi=120);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Important features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " 24 Important features(> 15 % of max importance):\n",
      " Features sorted by importance (DESC):\n",
      " [12 15 16 17 18 4 3 1 5 2 11 0 14 21 13 22 6 8 7 19 20 10 9 23]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlMAAAJcCAYAAADdFyE9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xa82d278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "MLalgo = regr_rf\n",
    "features_list = featuresActiv_list.copy()\n",
    "\n",
    "# make importances relative to max importance\n",
    "feature_importance = 100.0 * (MLalgo.feature_importances_ / MLalgo.feature_importances_.max())\n",
    " \n",
    "# A threshold below which to drop features from the final data set. Specifically, this number represents\n",
    "# the percentage of the most important feature's importance value\n",
    "fi_threshold = 15\n",
    " \n",
    "# Get the indexes of all features over the importance threshold\n",
    "important_idx = np.where(feature_importance > fi_threshold)[0]\n",
    " \n",
    "# Create a list of all the feature names above the importance threshold\n",
    "important_features = features_list[important_idx]\n",
    "#print (\"\\n\",important_features.shape[0],\"Important features(>\", fi_threshold, \"% of max importance):\\n\",\\important_features)\n",
    " \n",
    "# Get the sorted indexes of important features\n",
    "sorted_idx = np.argsort(feature_importance[important_idx])[::-1]\n",
    "print (\"\\n\", important_features.shape[0], \"Important features(>\", fi_threshold, \"% of max importance):\\n\",\n",
    "       \"Features sorted by importance (DESC):\\n\", important_features[sorted_idx])\n",
    " \n",
    "# Adapted from http://scikit-learn.org/stable/auto_examples/ensemble/plot_gradient_boosting_regression.html\n",
    "figsize(12,12)\n",
    "pos = np.arange(sorted_idx.shape[0]) + .5\n",
    "plt.rcParams[\"figure.figsize\"]=10,10\n",
    "plt.subplot(1, 1, 1)\n",
    "plt.barh(pos, feature_importance[important_idx][sorted_idx[::-1]], align='center')\n",
    "plt.yticks(pos, important_features[sorted_idx[::-1]])\n",
    "plt.xlabel('Relative Importance')\n",
    "plt.title('Variable Importance')\n",
    "plt.draw()\n",
    "#plt.show()\n",
    "\n",
    "# plt.savefig(\"ActivData01ImpFeatures.png\",dpi=120)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "###########################################\n",
    "\n",
    "### Tovabbi iranyok/feladatok:\n",
    "\n",
    "# kulonbozoen beallitott random forest osszekapcsolasa\n",
    "# \n",
    "# a ket modszer osszekapcsolasa\n",
    "# \n",
    "# ...\n",
    "\n",
    "###########################################"
   ]
  },
  {
   "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}