Titanic/notebooks/053-Titanic_new.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Another Submission - adding more features to Random Forest. \n",
    "\n",
    "All this is taken from https://www.kaggle.com/creepykoala/titanic/study-of-tree-and-forest-algorithms/notebook \n",
    "I just write it in line by line by myself as it helps me learn and I can query things as I go. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Libraries Ready!\n"
     ]
    }
   ],
   "source": [
    "# Import libraries\n",
    "\n",
    "import numpy as np\n",
    "from numpy.random import random_integers\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier\n",
    "from scipy.stats import pointbiserialr, spearmanr\n",
    "%matplotlib inline\n",
    "\n",
    "print('Libraries Ready!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Load training data\n",
    "\n",
    "df = pd.read_csv('/home/sophie/projects/Titanic/data/train.csv', header=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "People with stronger titles tend to have more help on board. Hence, we will categorize passengers based on titles."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0      Mr\n",
      "1     Mrs\n",
      "2    Miss\n",
      "3     Mrs\n",
      "4      Mr\n",
      "Name: Title, dtype: object\n",
      "0                              Braund, Mr. Owen Harris\n",
      "1    Cumings, Mrs. John Bradley (Florence Briggs Th...\n",
      "2                               Heikkinen, Miss. Laina\n",
      "3         Futrelle, Mrs. Jacques Heath (Lily May Peel)\n",
      "4                             Allen, Mr. William Henry\n",
      "Name: Name, dtype: object\n"
     ]
    }
   ],
   "source": [
    "Title_Dictionary = {\n",
    "                    \"Capt\":       \"Officer\",\n",
    "                    \"Col\":        \"Officer\",\n",
    "                    \"Major\":      \"Officer\",\n",
    "                    \"Jonkheer\":   \"Royalty\",\n",
    "                    \"Don\":        \"Royalty\",\n",
    "                    \"Sir\" :       \"Royalty\",\n",
    "                    \"Dr\":         \"Officer\",\n",
    "                    \"Rev\":        \"Officer\",\n",
    "                    \"the Countess\":\"Royalty\",\n",
    "                    \"Dona\":       \"Royalty\",\n",
    "                    \"Mme\":        \"Mrs\",\n",
    "                    \"Mlle\":       \"Miss\",\n",
    "                    \"Ms\":         \"Mrs\",\n",
    "                    \"Mr\" :        \"Mr\",\n",
    "                    \"Mrs\" :       \"Mrs\",\n",
    "                    \"Miss\" :      \"Miss\",\n",
    "                    \"Master\" :    \"Master\",\n",
    "                    \"Lady\" :      \"Royalty\"\n",
    "                    }\n",
    "\n",
    "# for each name, i.e. Braund, Mr. Owen Harris, we extract the title using x.split and strip and use it as a key to the \n",
    "# Title Dictionary. \n",
    "# x.split(',')[1] means split the string where the , is. Take the 2nd item. i.e. Mr. Owen Harris\n",
    "# split('.')[0] means split by the . and take the first item.\n",
    "# strip() to remove all the other characters. \n",
    "df['Title'] = df['Name'].apply(lambda x: Title_Dictionary[x.split(',')[1].split('.')[0].strip()])\n",
    "\n",
    "print(df['Title'][0:5])\n",
    "print(df['Name'][0:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['Futrelle,', 'Mrs.', 'Jacques', 'Heath', '(Lily', 'May', 'Peel)']]"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "name = ['Braund, Mr. Owen Harris', 'Futrelle, Mrs. Jacques Heath (Lily May Peel)']\n",
    "[name[1].split(',')[1].split('.')[0].strip()]\n",
    "[name[1].split()]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "Include the ticket prefix as a new column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    A/5 21171\n",
      "1     PC 17599\n",
      "Name: Ticket, dtype: object\n"
     ]
    }
   ],
   "source": [
    "print (df['Ticket'][0:2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0         A/5\n",
      "1          PC\n",
      "2    STON/O2.\n",
      "3      NoClue\n",
      "4      NoClue\n",
      "Name: TicketPrefix, dtype: object\n"
     ]
    }
   ],
   "source": [
    "# Extract the letters from the beginning of each element of 'Ticket'\n",
    "def Ticket_Prefix(s):\n",
    "    s=s.split()[0] # if you don't include anything in split()\n",
    "    if s.isdigit():\n",
    "        return 'NoClue'\n",
    "    else:\n",
    "        return s\n",
    "    \n",
    "df['TicketPrefix'] = df['Ticket'].apply(lambda x: Ticket_Prefix(x))\n",
    "\n",
    "print(df['TicketPrefix'][0:5])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fill in missing age with median age, based on Title, Pclass and Sex.\n",
    "\n",
    "- Make a mask with non-null values for Age. Make a new dataframe out of this new mask.\n",
    "- Make a groupby object from these values and calculate the median for each group"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Make an array where null values are False.\n",
    "mask_Age = df.Age.notnull()\n",
    "\n",
    "# New dataframe where all rows have a value for age. \n",
    "Age_Sex_Title_Pclass = df.loc[mask_Age, [\"Age\", \"Title\", \"Sex\", \"Pclass\"]]\n",
    "\n",
    "# Groupby object to group by Title, Pclass and Sex\n",
    "Filler_Ages_1 = Age_Sex_Title_Pclass.groupby(by = [\"Title\", \"Pclass\", \"Sex\"]).median()\n",
    "\n",
    "# This moves both Sex and Pclass into column headers and does so in that order. \n",
    "Filler_Ages = Filler_Ages_1.Age.unstack(level = -1).unstack(level = -1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Age Title     Sex  Pclass\n",
      "0  22.0    Mr    male       3\n",
      "1  38.0   Mrs  female       1\n",
      "2  26.0  Miss  female       3\n",
      "3  35.0   Mrs  female       1\n",
      "4  35.0    Mr    male       3\n",
      "714\n"
     ]
    }
   ],
   "source": [
    "print(Age_Sex_Title_Pclass[0:5])\n",
    "print(len(df[mask_Age]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Title     Sex  Pclass\n",
      "5     Mr    male       3\n",
      "17    Mr    male       2\n",
      "19   Mrs  female       3\n",
      "26    Mr    male       3\n",
      "28  Miss  female       3\n"
     ]
    }
   ],
   "source": [
    "mask_Age = df.Age.isnull()  # A mask where null values are True\n",
    "\n",
    "# New DataFrame with missing values for age\n",
    "Age_Sex_Title_Pclass_missing = df.loc[mask_Age, [\"Title\", \"Sex\", \"Pclass\"]]\n",
    "\n",
    "print(Age_Sex_Title_Pclass_missing[0:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Age Title     Sex  Pclass\n",
      "0  22.0    Mr    male       3\n",
      "1  38.0   Mrs  female       1\n",
      "2  26.0  Miss  female       3\n",
      "3  35.0   Mrs  female       1\n",
      "4  35.0    Mr    male       3\n",
      "Before unstacking---------\n",
      "                        Age\n",
      "Title   Pclass Sex         \n",
      "Master  1      male     4.0\n",
      "        2      male     1.0\n",
      "        3      male     4.0\n",
      "Miss    1      female  30.0\n",
      "        2      female  24.0\n",
      "        3      female  18.0\n",
      "Mr      1      male    40.0\n",
      "        2      male    31.0\n",
      "        3      male    26.0\n",
      "Mrs     1      female  40.0\n",
      "        2      female  31.5\n",
      "        3      female  31.0\n",
      "Officer 1      female  49.0\n",
      "               male    51.0\n",
      "        2      male    46.5\n",
      "Royalty 1      female  40.5\n",
      "               male    40.0\n",
      "After unstacking---------\n",
      "Sex     female              male            \n",
      "Pclass       1     2     3     1     2     3\n",
      "Title                                       \n",
      "Master     NaN   NaN   NaN   4.0   1.0   4.0\n",
      "Miss      30.0  24.0  18.0   NaN   NaN   NaN\n",
      "Mr         NaN   NaN   NaN  40.0  31.0  26.0\n",
      "Mrs       40.0  31.5  31.0   NaN   NaN   NaN\n",
      "Officer   49.0   NaN   NaN  51.0  46.5   NaN\n",
      "Royalty   40.5   NaN   NaN  40.0   NaN   NaN\n"
     ]
    }
   ],
   "source": [
    "print(Age_Sex_Title_Pclass[0:5])\n",
    "print(\"Before unstacking---------\")\n",
    "print (Filler_Ages_1)\n",
    "print(\"After unstacking---------\")\n",
    "print (Filler_Ages)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Make a function, Age_filler, which is applied to each of the missing dataframe."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Look-up function for the calculated median ages. \n",
    "def Age_filler(row):\n",
    "    if row.Sex == \"female\":\n",
    "        age = Filler_Ages.female.loc[row[\"Title\"], row[\"Pclass\"]]\n",
    "        return age\n",
    "    elif row.Sex == \"male\":\n",
    "        age = Filler_Ages.male.loc[row[\"Title\"], row[\"Pclass\"]]\n",
    "        return age\n",
    "    \n",
    "# Make a new column on \"missing\" dataframe and add the median value to each row. \n",
    "Age_Sex_Title_Pclass_missing[\"Age\"] = Age_Sex_Title_Pclass_missing.apply(Age_filler, axis= 1)\n",
    "\n",
    "# reform the 'Age' column.\n",
    "df[\"Age\"] = pd.concat([Age_Sex_Title_Pclass[\"Age\"], Age_Sex_Title_Pclass_missing[\"Age\"]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is one missing Fare value to fill in:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Filling in with the mean of all fares.\n",
    "df['Fare'] = df['Fare'].fillna(value=df.Fare.mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop Cabin and Ticket.   \n",
    "Combine SibSp and Parch to create FamilySize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df['FamilySize'] = df['SibSp'] + df['Parch']\n",
    "df = df.drop(['Ticket', 'Cabin'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will use dummy variables to transform the categorical data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0      male\n",
      "1    female\n",
      "2    female\n",
      "Name: Sex, dtype: object\n",
      "   Sex_female  Sex_male\n",
      "0         0.0       1.0\n",
      "1         1.0       0.0\n",
      "2         1.0       0.0\n"
     ]
    }
   ],
   "source": [
    "# get_dummies splits up a column into two seperate columns of 1 and 0, where they are true or false. \n",
    "dummies_Sex = pd.get_dummies(df['Sex'], prefix='Sex')\n",
    "\n",
    "print(df['Sex'][0:3])\n",
    "print(dummies_Sex[0:3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Making dummies for the other categorical features\n",
    "dummies_Embarked = pd.get_dummies(df['Embarked'], prefix= 'Embarked') \n",
    "dummies_Pclass = pd.get_dummies(df['Pclass'], prefix= 'Pclass')\n",
    "dummies_Title = pd.get_dummies(df['Title'], prefix= 'Title')\n",
    "dummies_TicketPrefix = pd.get_dummies(df['TicketPrefix'], prefix='TicketPrefix')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Make new dataframes which have the dummies added on to the end.\n",
    "df = pd.concat([df, dummies_Sex, dummies_Embarked, dummies_Pclass, dummies_Title, dummies_TicketPrefix], axis=1)\n",
    "\n",
    "# Drop the categorical data\n",
    "df = df.drop(['Sex','Embarked','Pclass','Title','Name','TicketPrefix'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, set PassengerId as the index:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Survived</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Fare</th>\n",
       "      <th>FamilySize</th>\n",
       "      <th>Sex_female</th>\n",
       "      <th>Sex_male</th>\n",
       "      <th>Embarked_C</th>\n",
       "      <th>Embarked_Q</th>\n",
       "      <th>...</th>\n",
       "      <th>TicketPrefix_SOTON/O.Q.</th>\n",
       "      <th>TicketPrefix_SOTON/O2</th>\n",
       "      <th>TicketPrefix_SOTON/OQ</th>\n",
       "      <th>TicketPrefix_STON/O</th>\n",
       "      <th>TicketPrefix_STON/O2.</th>\n",
       "      <th>TicketPrefix_SW/PP</th>\n",
       "      <th>TicketPrefix_W./C.</th>\n",
       "      <th>TicketPrefix_W.E.P.</th>\n",
       "      <th>TicketPrefix_W/C</th>\n",
       "      <th>TicketPrefix_WE/P</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PassengerId</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7.2500</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>71.2833</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.9250</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>53.1000</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0500</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 64 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             Survived   Age  SibSp  Parch     Fare  FamilySize  Sex_female  \\\n",
       "PassengerId                                                                  \n",
       "1                   0  22.0      1      0   7.2500           1         0.0   \n",
       "2                   1  38.0      1      0  71.2833           1         1.0   \n",
       "3                   1  26.0      0      0   7.9250           0         1.0   \n",
       "4                   1  35.0      1      0  53.1000           1         1.0   \n",
       "5                   0  35.0      0      0   8.0500           0         0.0   \n",
       "\n",
       "             Sex_male  Embarked_C  Embarked_Q        ...          \\\n",
       "PassengerId                                          ...           \n",
       "1                 1.0         0.0         0.0        ...           \n",
       "2                 0.0         1.0         0.0        ...           \n",
       "3                 0.0         0.0         0.0        ...           \n",
       "4                 0.0         0.0         0.0        ...           \n",
       "5                 1.0         0.0         0.0        ...           \n",
       "\n",
       "             TicketPrefix_SOTON/O.Q.  TicketPrefix_SOTON/O2  \\\n",
       "PassengerId                                                   \n",
       "1                                0.0                    0.0   \n",
       "2                                0.0                    0.0   \n",
       "3                                0.0                    0.0   \n",
       "4                                0.0                    0.0   \n",
       "5                                0.0                    0.0   \n",
       "\n",
       "             TicketPrefix_SOTON/OQ  TicketPrefix_STON/O  \\\n",
       "PassengerId                                               \n",
       "1                              0.0                  0.0   \n",
       "2                              0.0                  0.0   \n",
       "3                              0.0                  0.0   \n",
       "4                              0.0                  0.0   \n",
       "5                              0.0                  0.0   \n",
       "\n",
       "             TicketPrefix_STON/O2.  TicketPrefix_SW/PP  TicketPrefix_W./C.  \\\n",
       "PassengerId                                                                  \n",
       "1                              0.0                 0.0                 0.0   \n",
       "2                              0.0                 0.0                 0.0   \n",
       "3                              1.0                 0.0                 0.0   \n",
       "4                              0.0                 0.0                 0.0   \n",
       "5                              0.0                 0.0                 0.0   \n",
       "\n",
       "             TicketPrefix_W.E.P.  TicketPrefix_W/C  TicketPrefix_WE/P  \n",
       "PassengerId                                                            \n",
       "1                            0.0               0.0                0.0  \n",
       "2                            0.0               0.0                0.0  \n",
       "3                            0.0               0.0                0.0  \n",
       "4                            0.0               0.0                0.0  \n",
       "5                            0.0               0.0                0.0  \n",
       "\n",
       "[5 rows x 64 columns]"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.set_index(['PassengerId'])\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Feature Selection\n",
    "\n",
    "To select features we will correlate each feature against Survived. We need to use different algorithms for the different data types:\n",
    "\n",
    "- Spearman-Rank correlation for nominal vs nominal data\n",
    "- Point-Biserial correlation for nominal vs continuous data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Put the column names into a list called columns\n",
    "columns = df.columns.values\n",
    "\n",
    "# initiate lists to store results\n",
    "param = []\n",
    "correlation = []\n",
    "abs_corr=[]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "64\n",
      "Survived\n",
      "Age\n",
      "SibSp\n",
      "Parch\n",
      "Fare\n",
      "FamilySize\n",
      "Sex_female\n",
      "Sex_male\n",
      "Embarked_C\n",
      "Embarked_Q\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[None, None, None, None, None, None, None, None, None, None]"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(len(columns))\n",
    "#[print(type(df[c])) for c in columns][0:10]\n",
    "[print(\"%s\"%c) for c in columns[0:10]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for c in columns:\n",
    "    # Check if binary or continuous\n",
    "    if len(df[c].unique())<=2:\n",
    "        corr = spearmanr(df['Survived'], df[c])[0]   # apply spearmanr if\n",
    "    else:\n",
    "        corr = pointbiserialr(df['Survived'], df[c])[0]\n",
    "    \n",
    "    param.append(c) # name of feature\n",
    "    correlation.append(corr) # associated correlation\n",
    "    abs_corr.append(abs(corr))# get rid of -ve values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1.0, -0.062899950141403885, -0.035322498885735583, 0.081629407083483485, 0.25730652238496243]\n",
      "[1.0, 0.062899950141403885, 0.035322498885735583, 0.081629407083483485, 0.25730652238496243]\n"
     ]
    }
   ],
   "source": [
    "print(correlation[0:5])\n",
    "print(abs_corr[0:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "64\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>abs_corr</th>\n",
       "      <th>correlation</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>parameter</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Survived</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Title_Mr</th>\n",
       "      <td>0.549199</td>\n",
       "      <td>-0.549199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Sex_male</th>\n",
       "      <td>0.543351</td>\n",
       "      <td>-0.543351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Sex_female</th>\n",
       "      <td>0.543351</td>\n",
       "      <td>0.543351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Title_Mrs</th>\n",
       "      <td>0.344935</td>\n",
       "      <td>0.344935</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Title_Miss</th>\n",
       "      <td>0.332795</td>\n",
       "      <td>0.332795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Pclass_3</th>\n",
       "      <td>0.322308</td>\n",
       "      <td>-0.322308</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Pclass_1</th>\n",
       "      <td>0.285904</td>\n",
       "      <td>0.285904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fare</th>\n",
       "      <td>0.257307</td>\n",
       "      <td>0.257307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Embarked_C</th>\n",
       "      <td>0.168240</td>\n",
       "      <td>0.168240</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            abs_corr  correlation\n",
       "parameter                        \n",
       "Survived    1.000000     1.000000\n",
       "Title_Mr    0.549199    -0.549199\n",
       "Sex_male    0.543351    -0.543351\n",
       "Sex_female  0.543351     0.543351\n",
       "Title_Mrs   0.344935     0.344935\n",
       "Title_Miss  0.332795     0.332795\n",
       "Pclass_3    0.322308    -0.322308\n",
       "Pclass_1    0.285904     0.285904\n",
       "Fare        0.257307     0.257307\n",
       "Embarked_C  0.168240     0.168240"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create dataFrame for visualisation\n",
    "param_df = pd.DataFrame({'correlation':correlation, 'parameter':param, 'abs_corr':abs_corr})\n",
    "\n",
    "# Sort by absolute correlation\n",
    "param_df = param_df.sort_values(by=['abs_corr'], ascending=False)\n",
    "\n",
    "# Set parameter name as index\n",
    "param_df = param_df.set_index('parameter')\n",
    "\n",
    "print(len(param_df))\n",
    "param_df[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Survived                   0\n",
      "Age                        0\n",
      "SibSp                      0\n",
      "Parch                      0\n",
      "Fare                       0\n",
      "FamilySize                 0\n",
      "Sex_female                 0\n",
      "Sex_male                   0\n",
      "Embarked_C                 0\n",
      "Embarked_Q                 0\n",
      "Embarked_S                 0\n",
      "Pclass_1                   0\n",
      "Pclass_2                   0\n",
      "Pclass_3                   0\n",
      "Title_Master               0\n",
      "Title_Miss                 0\n",
      "Title_Mr                   0\n",
      "Title_Mrs                  0\n",
      "Title_Officer              0\n",
      "Title_Royalty              0\n",
      "TicketPrefix_A./5.         0\n",
      "TicketPrefix_A.5.          0\n",
      "TicketPrefix_A/4           0\n",
      "TicketPrefix_A/4.          0\n",
      "TicketPrefix_A/5           0\n",
      "TicketPrefix_A/5.          0\n",
      "TicketPrefix_A/S           0\n",
      "TicketPrefix_A4.           0\n",
      "TicketPrefix_C             0\n",
      "TicketPrefix_C.A.          0\n",
      "                          ..\n",
      "TicketPrefix_F.C.C.        0\n",
      "TicketPrefix_Fa            0\n",
      "TicketPrefix_LINE          0\n",
      "TicketPrefix_NoClue        0\n",
      "TicketPrefix_P/PP          0\n",
      "TicketPrefix_PC            0\n",
      "TicketPrefix_PP            0\n",
      "TicketPrefix_S.C./A.4.     0\n",
      "TicketPrefix_S.C./PARIS    0\n",
      "TicketPrefix_S.O./P.P.     0\n",
      "TicketPrefix_S.O.C.        0\n",
      "TicketPrefix_S.O.P.        0\n",
      "TicketPrefix_S.P.          0\n",
      "TicketPrefix_S.W./PP       0\n",
      "TicketPrefix_SC            0\n",
      "TicketPrefix_SC/AH         0\n",
      "TicketPrefix_SC/PARIS      0\n",
      "TicketPrefix_SC/Paris      0\n",
      "TicketPrefix_SCO/W         0\n",
      "TicketPrefix_SO/C          0\n",
      "TicketPrefix_SOTON/O.Q.    0\n",
      "TicketPrefix_SOTON/O2      0\n",
      "TicketPrefix_SOTON/OQ      0\n",
      "TicketPrefix_STON/O        0\n",
      "TicketPrefix_STON/O2.      0\n",
      "TicketPrefix_SW/PP         0\n",
      "TicketPrefix_W./C.         0\n",
      "TicketPrefix_W.E.P.        0\n",
      "TicketPrefix_W/C           0\n",
      "TicketPrefix_WE/P          0\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(df.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have our correlation, we can use the Decision Tree classifier to see the score against feature space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "scoresCV = []\n",
    "scores = []\n",
    "\n",
    "# Loop over each of the 64 different parameters\n",
    "for i in range(1, len(param_df)):\n",
    "    \n",
    "    # Make a new dataframe for each feature(selected by df[])\n",
    "    new_df = df[param_df.index[0:i+1].values]\n",
    "    \n",
    "    # .ix supports mixed labels and integer. It can do what .loc and .iloc. both do.\n",
    "    # X is abs_corr\n",
    "    X = new_df.ix[:,1::]\n",
    "    #print(X)\n",
    "    \n",
    "    # y is corr\n",
    "    y = new_df.ix[:,0]\n",
    "    \n",
    "    # initialise a DecisionTreeClassifier object\n",
    "    clf = DecisionTreeClassifier()\n",
    "    \n",
    "    # cross_val_score, returns: Array of scores of the estimator for each run of the cross validation.\n",
    "    # (estimators, data-to-fit, target-variable-to-predict, cv(determines the cross-validation splitting strategy))\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv = 10)\n",
    "    \n",
    "    # save the mean of all the runs\n",
    "    scores.append(np.mean(scoreCV))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Survived']\n",
      "['Survived', 'Title_Mr', 'Sex_male']\n",
      "63\n"
     ]
    }
   ],
   "source": [
    "new_df = df[param_df.index[0:1].values]\n",
    "print(list(new_df))\n",
    "new_df = df[param_df.index[0:3].values]\n",
    "print(list(new_df))\n",
    "print(len(scores))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot the scores for different numbers of features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA38AAAFVCAYAAABfH1b2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XeYFFXWBvD3EEUYRVQQUMSMzKiwfiIg4ogJs2ICEUVd\nxJwwZ1cRFQNmUNFVFDCusq6gogwosArCkJMKqATJ0OQJ5/vjdO80w9RMh+ququ739zzzMNVdXXVn\nuNNdp+6554qqgoiIiIiIiDJbNa8bQERERERERKnH4I+IiIiIiCgLMPgjIiIiIiLKAgz+iIiIiIiI\nsgCDPyIiIiIioizA4I+IiIiIiCgLMPgjIiLyCRF5WERmpOE8Y0TkxVSfh4iI/IXBHxERpYyIvC0i\npSJSEv438v2RLp4jLQFTDO04QURGi8hKEdkkIr+IyBARqRfnoVxbgFdErhCRUAVPnQ/gXrfOQ0RE\nwcDgj4iIUu0bAPtEfTUGMNPlc7gSMIlIzQRfdziAkQAKAeQDyAXQG8B6ALXdaFuCBBX8blR1napu\n8qA9RETkIQZ/RESUattUdaWqroj6Ko08KSJ3hUfJNovINBHpHv1iEeknInPDzy8UkadEpFb4uSsA\nPAwgN2pU8fLwc6Ui0qXcsRaKyO1R26Uicr2IfCIiGwH0DT/eUkS+EJENIvKXiAwVkUaV/IynAlil\nqneo6ixVXaSq36rqjaq6Oup88R4XInKliMwSkS3h38Ot5Z7fTUReE5Gl4X1michFInICgLcA1I36\n3TwUfs0OaZ8iUl9E3hGRNeHf8zci0jLq+StEJCQinURkhohsFJHvRGT/ytpORET+wuCPiIg8IyJ9\nAVwJ4DoAhwPoB2CgiJwetdtGAD0BtAjvdwmA+8PPfQDgWQDzADSCjSp+EGczHgLwHwB5AF4RkX0A\njAUwHcD/ATgJQF0An1dyjOUA9haRfKcdEjmuiPQC8DiAB2A/fx8Ad4nI9VG7jQRwPIArwvvcAmAb\ngPEAbgWwGWW/m2ccTvUOgGMAnB3+dzOAUSISPWpZG8A9sP+LtgDqAxjo1HYiIvKfGl43gIiIMt7p\n5eadjVPVM0VkVwC3AThFVceHn1ssIscCuAEW1EBV+0a99ncR6QcLgh5W1a3hEbtiVV2ZYPuGq+pb\nkQ0ReRRAoareF/VYTwCrReT/VHVyBcf4CDb6962IrATwE4AxAIao6qrwPtclcNwHANylqv8Kby8W\nkadgv59XReQUAMcCaKmq8yP7RB1/PQCt7HcjIofAgr7jI/8PItIDwO8AusNGDwGgOoDrVfWX8D7P\nABjsdFwiIvIfBn9ERJRqYwH0gs0/A4At4X9bAtgFNsIUvX8NAAsjGyJyIWw062AA9WBBiJuZKz+X\n2z4awAkVFEpRAAcB2ClIC6exXi0iDwDoBBsZuxPA/SJyvKrOife4IrIXgP0ADBKR6BG2Giibx9cK\nwLKowC8RLQCUAPhv1M+zIVxEp2XUftsigV/YUgC1RKS+qq5L4vxERJQmDP6IiCjVNqvqwgoejwRw\nZwH4o9xzRQAgIm0BDIPN6/sKwDoA5wLoH8N5FWUBZ0RFBV3KFz6pBuAL2Ohi+df/VekJVZcBeB/A\n++FAcAEsCLwqgeNGfj+9AUys7LwpFF0sptjhOU4hISIKCAZ/RETkldmwuWnNVXWswz7tAfypqk9E\nHhCR5uX22Q4bDSxvJWyeW+R1jaK3KzEFwEUAflfVkhj2r5CqrheRZbDRyriPq6orRGQpgINV9X2H\n3aYCaCwih6nqvAqed/rdRJsDC+DaAfgBsCIyAI4A0zqJiDIK79YREZEnVHUjrADJM+GKlgeJyFEi\n0ltE/h7ebT6ApiJyqYgcICLXAeha7lCLAOwvIq1FZM9IJVAA3wG4QUSOFpHWAN5GWcppZV4BsDuA\nD0WkTfi8J4vIIBGpW9ELROQaEXlVRE4RkQPDVT2fghWR+TTR48JGPO8SkVtF5FARyRWRHiJyT/j5\nb2HzCz8RkVNFpHn4mOdG/W52CT+2p4jUKX+CcCrnCFh6aQcROQLAe7BlKoZV8bsqP4JJREQ+xuCP\niIg8o6oPAngElgo5E8DXALogPOdPVb+ApXg+D2AarELmg+UO8wmAL2GB0AqUBYd9APwGK7zyIYA3\nws/v0IQK2rQMwHGweXAjw+16CcBW2EhlRX4CUAfAqwBmwOY5Hg+gh6oOT/S4qjoYljJ6GWwNwXGw\n+ZO/hZ9XAJ1hlT2HwEZTByCc3qqqE2EVOYeFf/Y7HdrfM/wzfA6b+1cbQGdVdfp5/9fEKp4nIiIf\nEfvcICIiIiIiokzGkT8iIiIiIqIswOCPiIiIiIgoCzD4IyIiIiIiygIM/oiIiIiIiLJAxq3zJyKs\nYENERERERFlNVXdajicjR/5UlV8Z/PXwww973gavv0pLFa1bK5o2VQCKFi0UGzZ43y6vv9g3+OX0\nxb7BL6cv9g1+OX2xb/DL6SsIfcNJRgZ/lNkWLVrkdRM8N3o0sG0bMGsWcN11wIEHAjk5XrfKe+wb\n5IR9g5ywb5AT9g1yEuS+weCPKICeeAK4915g992B554DZs8GvvvO61YRERERkZ8x+KPA6dmzp9dN\n8NSECcDixUDXrra9yy5A//7AbbcBJSXets1r2d43yBn7Bjlh3yAn7BvkJMh9QyrLCQ0iEdFM+5mI\nop19NnDmmcC115Y9pgqccALQowfQq5d3bSMiIiIi74kINFsKvlBmKygo8LoJnpk2Dfj5Z6D8DScR\n4PnngYceAjZs8KRpvpDNfYMqx75BTtg3yAn7BjkJct9g8EcUIP36Abffbqme5R19NNC5s80HJCIi\nIiIqj2mfRAExfz5w3HHAb785V/ZcuhQ44ghg0iSrAEpERERE2Ydpn0QB9/TTwA03VL6kQ5MmVvjl\n7rvT1y4iIiIiCgYGfxQ4Qc6zTtQffwCffgrcdFPV+/bpA/z0EzBuXOrb5TfZ2DcoNuwb5IR9g5yw\nb5CTIPcNBn9EAfDss8DVVwN77ln1vnXqAE89ZSOApaWpbxsRERERBQPn/BH53MqVwGGHAbNmAY0b\nx/YaVZsfeM01O1cGJSIiIqLM5jTnj8Efkc898ACwejXw2mvxve7HH4EuXYB584B69VLTNiIiIiLy\nHxZ8oYwR5DzreK1fDwwcCNx5Z/yvPfZYoFMn4Mkn3W+XX2VT36D4sG+QE/YNcsK+QU6C3DcY/BH5\n2KuvAqefnviyDf362Yjh4sXutouIiIiIgodpn0Q+tXmzBX3ffgvk5iZ+nEcesdTPYcNcaxoRERER\n+RjTPonSJBQCJk60f5MxeDDQrl1ygR9gKaM//ABMmJDccYiIiIgo2Bj8UeD4Oc96wwZgv/2A44+3\nr0QDwO3bgf79gfvuS75NdesCTzyRHUs/+LlvkLfYN8gJ+wY5Yd8gJ0HuGwz+iFz02WdWpKWkBJgx\nA5g8ObHjvP8+0KIFcMwx7rSre3db/mHoUHeOR0RERETBwzl/RC565BHg5ZdtBLBuXVuUfehQoE2b\n2I9RUgK0bAkMGgTk57vXtgkTgEsuAebOtbYRERERUWbyxZw/EeksInNFZL6I3F3B87uJyAgRKRSR\nGSLSM/z4viLynYjMCj9+czrbTRSr776zoG3cOOD3363a5tlnA48+ChQXx3aMTz+1oPGEE9xtW/v2\ntvD7M8+4e1wiIiIiCoa0BX8iUg3AywBOA5ALoJuItCi32w0AZqlqKwAnAnhWRGoAKAZwu6rmAmgH\n4IYKXktZwq951uvWAVOnAmecAbRtC+TkABddZI9NmGCB14IFlR9DFejb1+b6yU73apL31FPACy8A\nn3+efEEaP/Jr3yDvsW+QE/YNcsK+QU6C3DfSOfLXBsACVV2sqkUAhgM4t9w+CiAn/H0OgNWqWqyq\ny1W1EABUdSOAOQCapqndRDH55hugQwegTp0dH2/SBBg1CujRw0bfBg2yIK8iI0fac2eemZo2NmgA\n1KgBdOmSXEEaIiIiIgqetM35E5ELAJymqteEty8D0EZVb47apx6AEQBaAKgH4BJVHVnuOM0BFADI\nCweC5c/DOX/kiauuAlq3Bm66yXmfuXOByy4DGjWypRz22afsOVULHm+6CejaNTVtnDjRgr6SEqBm\nTUtPbds2NeciIiIiIm84zfmr4UVjKnEagKmq2klEDgLwjYgcGQnywsHhxwBuqSjwi+jZsyeaN28O\nAKhfvz5atWqF/HDljMgwLbe57eb2CSfkY+RI4KSTClBQ4Lz/8uUF6NcPGDs2H61aATfeWIAOHez5\n778HFi0qwN57A0Bq2rt2bQEaNQKWLs1Hy5bAmjWVt5fb3OY2t7nNbW5zm9v+3y4sLMS6desAAIsW\nLYKTdI78tQXwiKp2Dm/fA0BV9amofb4A0E9Vx4e3vwVwt6pODs/9+wLASFV9oZLzcOQvwxUUFPyv\ns/vF1KnAxRdXPacv2oQJwOWXAx07Ao8/bqmYPXoAN9yQunYCQGEhcPrpwPz5Ni8xk/ixb5A/sG+Q\nE/YNcsK+QU6C0Df8UO1zEoCDRWR/EakFoCssxTPaYgAnA4CINAJwKIDfws+9BWB2ZYEfkVdGjrRC\nL/Fo394CsZISYP/9gR9/tPmAqZ6Hd8ghwNq1QL16qT0PEREREflLWtf5E5HOAF6ABZ2DVfVJEekN\nGwF8XUQaA/gngMbhl/RT1WEichyAcQBmwIrCKID7VHVUBefgyB+l3fHHA/ffD3TuHP9rvZiHt/vu\nwMKFVgCGiIiIiDKL08gfF3knStLatTZy99dfO1f6jEUoZMHf7Nm2uPv336c+HbNFC+CTT4Dc3NSe\nh4iIiIjSzw9pn0SuiExy9YtvvrHgLZHAD7BA7/vvbcQvHYEfYMtPLFuW+vOkm9/6BvkH+wY5Yd8g\nJ+wb5CTIfcNv1T6JAmfkSCugkoycnPQuudC4MbB0afrOR0RERETeY9onURJKS20UbcIE4MADvW5N\n7O64A9h7b+Duu71uCRERERG5jWmfRClQWGjFU4IU+AE28peJaZ9ERERE5IzBHwWOn/Ks3Uj59ALn\n/FG2Yd8gJ+wb5IR9g5wEuW8w+CNKwpdfxr++nx9wzh8RERFR9uGcP6IErVkDNG8OrFgB7LKL162J\nz7x5wJlnAr/84nVLiIiIiMhtnPNH5LJvvgE6dgxe4AeUzfnjfRIiIiKi7MHgjwLHL3nWQU35BGxp\nCRFbYD6T+KVvkP+wb5AT9g1ywr5BToLcNxj8ESWgtBQYNSqYxV4AC/w474+IiIgou3DOHyUtFAJm\nzgTy8mxEKRv8/DPQvTswd67XLUlcx47Ao48CJ57odUuIiIiIyE2c80cpEQoBxx1ngcTxx2deGqGT\nIKd8RnCtPyIiIqLswuCPkjJzJjBrFlBcDMyebd+nmh/yrIO6vl+0Jk0yL+3TD32D/Il9g5ywb5AT\n9g1yEuS+weCPkpKXB+y9t32/225Abq637UmH1ast6O3Y0euWJIcjf0RERETZhXP+KGk33ABs2waM\nHg0MGACcd57XLUqtYcPsa8QIr1uSnCFDbARz6FCvW0JEREREbuKcP0qZLVuAdu0sIOrdG1iyxOsW\npVYmpHwCHPkjIiIiyjYM/ihpGzZYyme7dsCNNwKXX25LIaSKl3nWQV/iIRrn/FE2Yd8gJ+wb5IR9\ng5wEuW8w+KOkrV8P7L67fX/ffUBREfDMM962KVV+/hnYay+geXOvW5I8jvwRERERZRfO+aOktWkD\nvPQScOyxtv3778D//Z8th/B//+dt29z2j39YsPvss163JHmqQJ06VsCmbl2vW0NEREREbuGcP0qZ\nSNpnRLNmwMsvA5deCmzc6F27UmHkyOCv7xchwtE/IiIiomzC4I+SFp32GXHxxbb4+623un8+r/Ks\nV62ytQw7dPDk9CmRafP+gpyDT6nFvkFO2DfICfsGOQly32DwR0krP/IX8eKLwNixwEcfpb9NqfD1\n10B+PlC7ttctcQ9H/oiIiIiyB+f8UVKKimzeWFGRpRGWN2kScOaZwOTJlg4aZD162Khf795et8Q9\nN90EHHRQakZoiYiIiMgbnPNHKREZ9aso8AOAY44Bbr/dAqeSkvS2zU2ZtMRDtCZNOPJHRERElC0Y\n/FFSnFI+o915J1C9OvDkk+6c04s868mTgYYNgz96WV7jxpzzR9mBfYOcsG+QE/YNchLkvsHgj5JS\nUbGX8qpXB9591+YA/vhjetrltkyq8hmNc/6IiIiIsgfn/FFSxo4FHngA+P77qvf99FMbBZw6terR\nQr859ligXz+gUyevW+Ku6dOBbt2AWbO8bgkRERERuYVz/iglNmyoeuQvoksX4KSTrMhIkKxcCcyd\nm1lLPERwzh8RERFR9mDwR0mJJe0z2vPPW+rn0KGJnzPdedZff20jfrVqpfW0abHnnsCmTcCWLV63\nxB1BzsGn1GLfICfsG+SEfYOcBLlv1PC6ARRssRR8iVa3LjBsGHDaacARRwAbNwJ5eUBOTuramKwv\nv8y8Kp8RIkCjRsDy5cABB3jdGiIiIiJKJc75o6T062ejf/FW8uzbF3jiCWD7diA31+YM+jEALCkB\n9tkHmDIF2G8/r1uTGsceayOy7dt73RIiIiIicgPn/FFKxJv2GZGfD2zdChQXA9OmAffcAyxa5Hbr\nkjd5sgV/mRr4AZz3R0RERJQtGPxRUuJN+4w48khL+6xZEzjoIGDzZlsQvl074IUXKl97Lp151pmc\n8hmRSWv9BTkHn1KLfYOcsG+QE/YNchLkvpHW4E9EOovIXBGZLyJ3V/D8biIyQkQKRWSGiPSMem6w\niPwlItPT2WaqXKIjfzk5luo5bpwt/fD22xaAPPywbefmAieeCAwaBKxa5X67YzVyZHYEfxz5IyIi\nIsp8aZvzJyLVAMwHcBKApQAmAeiqqnOj9rkXwG6qeq+I7AVgHoBGqlosIh0AbATwrqoeWcl5OOcv\njc4+G+jVCzjnHHePu3UrMGoU8MEHNvrWrh3QtastFfHnn/EXiQmFgJkz43vdb7/ZCOXixVYVM1O9\n+SYwfrwF4EREREQUfE5z/tJZ7bMNgAWqujjcoOEAzgUwN2ofBRC5NM8BsFpViwFAVX8Qkf3T2F6K\nQaIjf1XZZRfgvPPsa9Mm4IsvgCFDgKuvBkpLgTp1gMMPt7TRatWcv0Rs/x9/tMqi9epZYZMaVfT8\n4mIbldy61QJOvxakcQPn/BERERFlh3QGf00B/BG1/ScsIIz2MoARIrIUQD0Al6SpbZSgVAV/0erW\nBS65BGjWDPjqK6C0tABFRfm4/npLDy0trfxr9mxgzBhA1dazO/lkoEWLys85dy4werS9ZvZsYNYs\noG3b1P6cXsm0OX/5+fleN4N8iH2DnLBvkBP2DXIS5L7ht3X+TgMwVVU7ichBAL4RkSNVdWM8B+nZ\nsyeaN28OAKhfvz5atWr1v/+gyARNbruz/ddfBZg9G2jVKvXny8sD9t+/AAsXFiI3Nx8XXwz8/PPO\n+1erBnTqVLbdogWQl5eP2bOBZs0K0LIlcMYZlZ+vd+98vPceMHNmAZo1A3JzU//zebW9Zg2wbJl/\n2pPMdmFhoa/aw21uc9v/2xF+aQ+3/bNdWFjoq/Zw2z/bfrzeKCwsxLp16wAAiyopoZ/OOX9tATyi\nqp3D2/cAUFV9KmqfLwD0U9Xx4e1vAdytqpPD2/sD+Dfn/PlHgwbAggXpmxMXCtkoXG5u/HP+4n1d\noucKmpISS7PdtAmoVcvr1hARERFRsvywzt8kAAeLyP4iUgtAVwAjyu2zGMDJACAijQAcCuC3qOcl\n/EU+oJr4Ug+Jysmx9Mt4g7FEXpfouYKmenWgYUPgr7+8bgkRERERpVLagj9VLQFwI4CvAcwCMFxV\n54hIbxG5Jrzb4wDah5dz+AbAXaq6BgBEZCiACQAOFZHfReTKdLWdKrZ5s40U1ayZ3vNGhrrJPZky\n7499g5ywb5AT9g1ywr5BToLcN9I6509VRwE4rNxjg6K+Xwab91fRay9NbesoXuko9kLpwbX+iIiI\niDJf2ub8pQvn/KXP3LnAuecC8+Z53RJK1jXXAK1bA9dd53VLiIiIiChZfpjzRxmGI3+Zg2v9ERER\nEWU+Bn+UsHQXe4kIcp61X3HOH2U69g1ywr5BTtg3yEmQ+waDP0oYR/4yB+f8EREREWU+zvmjhA0e\nDIwfD7z1ltctoWRNnmzz/qZM8bolRERERJQszvkj13mV9knu48gfERERUeZj8EcJ8yrtM8h51n7V\nqBGwahVQXOx1S5LDvkFO2DfICfsGOWHfICdB7hsM/ihh69dz5C9T1KgBNGgArFjhdUuIiIiIKFU4\n548SdvXVQLt2wN//7nVLyA2tWwNvvgkcfbTXLSEiIiKiZHDOH7mO1T4zC+f9EREREWU2Bn+UMK7z\nl1kyYa0/9g1ywr5BTtg3yAn7BjkJct9g8EcJ48hfZuHIHxEREVFm45w/SliLFsCnnwItW3rdEnLD\nq68C06cDAwd63RIiIiIiSgbn/JHrNmzgyF8m4cgfERERUWZj8EcJ4zp/mYVz/iiTsW+QE/YNcsK+\nQU6C3DcY/FFCiouBrVuBunW9bgm5hSN/RERERJmNc/4oIWvWAAcdBKxd63VLyC3btgE5ORbUV+Nt\nISIiIqLA4pw/chUrfWae2rVt6Y5Vq7xuCRERERGlAoM/SohXa/wBwc6z9rugz/tj3yAn7BvkhH2D\nnLBvkJMg9w0Gf5QQjvxlJs77IyIiIspcDP4oIV6O/OXn53tz4izQpEmwgz/2DXLCvkFO2DfICfsG\nVSQUAmrXzkco5HVLEsPgjxLCkb/MFPS0TyIiIqJUCYWAdu2Ajh2B449HIANABn+UEC+DvyDnWftd\n0NM+2TfICfsGOWHfICfsG1TezJnAnDlAcXEBZs8GZs3yukXxY/BHCfEy7ZNSJ+hpn0RERESpsu++\nZd8fdhiQm+tdWxLFdf4oIffea2vC3Xef1y0hN40fD9xxBzBxotctISIiIvKXF18Exo2zKTK9egFX\nXul1i5xxnT9yFUf+MhPn/BERERHtTBUYOBC4+WbgqquAkSO9blFiGPxRQjjnLzM1bgwsX25vcEHE\nvkFO2DfICfsGOWHfoGjjxgEiVuhlzz0L8PXXwNatXrcqfgz+KCGs9pmZ6tQBdt0VWLPG65YQERER\n+cdrrwHXXmsB4B57AEcdBYwe7XWr4sc5f5SQE04AHn0U4BI4madlS+DDD4G8PK9bQkREROS9v/4C\nWrQAFi0qG/wYMACYMQMYPNjTpjninD9yFUf+Mhfn/RERERGVeest4IILdrz2Pe88YMQIoLjYu3Yl\ngsEfJcTLgi/MwU+tIK/1x75BTtg3yAn7Bjlh3yAAKCkBBg0Crruu7LGCggI0bw7stx/www+eNS0h\nDP4oIRz5y1xc64+IiIjIfPUVsPfewNFH7/xcly7Av/6V/jYlg3P+KG6qQM2awObNQK1aXreG3Pb8\n85bT/sILXreEiIiIyFtnnw2cf74t71DerFnA6acDixdbIRg/8cWcPxHpLCJzRWS+iNxdwfO7icgI\nESkUkRki0jPW11L6bNliwR8Dv8zEOX9EREREFtRNmAB07Vrx8y1bArvsAvz8c3rblYy0BX8iUg3A\nywBOA5ALoJuItCi32w0AZqlqKwAnAnhWRGrE+FpKE69TPpmDn1qc80eZiH2DnLBvkBP2DXrjDeCy\ny2wZrGiRviFio4JBSv1M58hfGwALVHWxqhYBGA7g3HL7KICc8Pc5AFaranGMr6U0Wb/eu2IvlHqc\n80dERETZbvt2W8bh2msr3y9o8/7SGfw1BfBH1Paf4ceivQygpYgsBTANwC1xvJbSZMMGb0f+8rm4\nYEpFRv6COHWWfYOcsG+QE/YNcsK+kd0++8zW9jv88J2fi+4bxxxjAyNz56avbcnwW7XP0wBMVdUm\nAFoDeEVE6nncJirH67RPSq169YDq1e3/mYiIiCgbDRxY9agfAFSrZmv+BWX0r0Yaz7UEQLOo7X3D\nj0W7EkA/AFDVX0VkIYAWMb72f3r27InmzZsDAOrXr49WrVr9L0KP5OhyO/HtCROA3Xbz7vyFhYW4\n9dZbffP7yMTtxo3zsWwZUFjoj/bEuj1gwAD+vXO7wu3I935pD7f9sx15zC/t4bZ/tnm9kb3b775b\ngMJC4PzzK36+/PXGgQcW4I03gHvv9a79hYWFWLduHQBg0aJFcJK2pR5EpDqAeQBOArAMwE8Auqnq\nnKh9XgGwQlUfFZFGACYDOArA+qpeG3UMLvWQYm+9BXz/PfD2296cv6Cg4H+dnVIjPx946CGgUyev\nWxIf9g1ywr5BTtg3yAn7Rva69Vagbl2gb9+Kny/fN4qKgH32AaZNA/bdNz1trIrTUg9pXedPRDoD\neAGWbjpYVZ8Ukd4AVFVfF5HGAP4JoHH4Jf1UdZjTax3OweAvxZ5/3krfDhjgdUsoVbp1A846C+je\n3euWEBEREaXP5s3AfvvZ8g3hRMKYXHGFzf+78caUNS0uTsFfOtM+oaqjABxW7rFBUd8vg837i+m1\n5I0NG1jtM9NxrT8iIiLKRh98ALRrF1/gB9iSDy+95J/gz0k1rxtAweN1wZdInjOlTlDX+mPfICfs\nG+SEfYOcsG9kp4EDgeuuq3yfivrGqacCkycDq1enpl1uYfBHcePIX+bjWn9ERESUbaZMAf76C+jc\nOf7X7rorcNJJwL//7X673MTgj+Lm9cgfJ1+nXlBH/tg3yAn7Bjlh3yAn7BvZ57XXgGuusSWvKuPU\nN9xe8D0UAiZOtH/dwuCP4uZ18Eepxzl/RERElE3Wrwc+/hi4+urEj3HmmcCYMcDGjcm3JxQC8vKA\nDh2A4493LwBk8Edx8zrtkzn4qRfUtE/2DXLCvkFO2DfICftGdhkyxObtNWpU9b5OfWOPPYC2bYFR\no5Jvz5dfAr//DpSWArNnA7NmJX9MgMEfJYAjf5lvt92AkhJ37lwRERER+ZmqpXxWVeglFuefn3zq\n5/btwJNPAk2bAtWqAQ0aALm5ybcNSPM6f+nAdf5Sr2lT4Mcf/bOIJaXGQQfZnatDDvG6JURERESp\n8/33Ntec/wbQAAAgAElEQVRv9mxAdloZLz7Lllmgtnw5UKtWYse4915gxgxg6FBg5EjghhuARYuA\nevViP4bTOn8c+aO4ceQvO3DeHxEREWWD114Drr02+cAPsOunFi1s7l8iCgqAd94B3nrLMrEuuQTo\n1Al4/fXk2wYw+KM4FRcDW7bEd+fBbczBT48gzvtj3yAn7BvkhH2DnLBvZIcVK2x07fLLY39NVX3j\n/POBTz+Nvy1r11o7Bg8GGjYse/zee4FnnwW2bo3/mOUx+KO4hEJATo47d0bI34K63AMRERFRrN5+\n24K1PfZw75jnnw98/rnVT4iVKtC7N3DeecDpp+/4XOvWQKtWNiKYLM75o7gsWgSccAKweLHXLaFU\ne/JJYM0a4OmnvW4JERERkftKS4GDDwY++AA45hh3j33EEcDAgcBxx8W2/zvvAP37A5MmAXXq7Pz8\n+PFAjx7A/PlAjRpVH49z/sgVXi/zQOnDOX9ERESUyT77zIqytGjh/rHjWfD911+BO+6wAi8VBX6A\nBZHNmgHDhiXXrpiDPxGpLSJ9ReQ3EVkffuxUEbkxuSZQkPih2Atz8NODc/4ok7BvkBP2DXLCvpHZ\nQiHgyiuBBQviX0Q9lr4RmfdXVUJiURHQvTtw//3AkUdWvu/99wP9+tmIZaLiGfl7HkAegO4AIj/G\nLAAurIhBQeGH4I/Sg3P+iIiIKFNNmWIZbW4voh5x1FEW+E2fXvl+jz9u19Y331z1MU8+Gahb10Ys\nExXznD8RWQbgYFXdJCJrVLVB+PF1qlo/8Sa4i3P+UmvoUODf/05+yJn8b/Vqy4Nfu9brlhARERG5\nq6DACquUlAAtW9pafzk57p7j9tttutQjj1T8/PjxwAUXAFOn2k33WHz2mQWMkyZVXoDRjTl/2wHs\nML1QRPYGsDqOY1DAceQvezRoAGzebEt7EBEREWWSOXOAiy4Cxo1LTeAHVD7vb/164LLLgEGDYg/8\nAOCcc2zJh6+/TqxN8QR/HwF4R0QOAAARaQzgZQDDEzs1BdH69d4XfGEOfnqIBC/1k32DnLBvkBP2\nDXLCvpHZJk8G2rcH2raNP/CLtW+0awcsX24FXcq78UbgtNOAc8+N79zVqtm6f337xve6/70+jn3v\nA7AQwAwA9QEsALAUwKOJnZqCaMMGjvxlk6AFf0RERESxmDwZ+L//S+05qle34K786N/QoZa2+eyz\niR33kkuAJUtsxDJeMc35E5FqAPIBjFfVbeF0z1V+nFzHOX+pdcMNwOGH290KynxdugCXXgpceKHX\nLSEiIiJyx+bNwF57WV2D2rVTe66RI22O3vjxtr1oka0p+NVXwN/+lvhxX3/dgsqRIyt+Pqk5f6pa\nCuBzVd0W3l7JCCs7cZ2/7MK1/oiIiCjTTJtmRV5SHfgBQKdOVk102TIrLtOjB3DnnckFfgBwxRXA\njBlWtTQe8aR9jhORtvEdnjKNHwq+MAc/fYK21h/7Bjlh3yAn7BvkhH0jcyWb8hlP36hd26qKfv45\n8OSTQM2atqB7smrXBvr0AZ54Ir7X1ah6l/9ZDGCkiHwO4A+UrfUHVX0ovtNSUPmh4AulT+PGVgWL\niIiIKFNMnmwLu6fL+efbAu0rVwITJljRFjdcc40FlHPm2LSsWMSzzt/bDk+pql4V2+lSj3P+Uqt1\na2Dw4OSHqikYRo4EBgywvHQiIiKiTJCbC7z/PtCqVXrOt2wZ0LSpBX15ee4uLdG3LzB/PvDOOzs+\n7jTnL+aRP1W9MvnmUdD5Ie2T0odz/oiIiCiTbNxoRVdyc9N3zkWLrPJncbHN/5s1y5aYcMMNNwAH\nHQQsXAgccEDV+8c16Cgih4jIQyIyKPzvIYk2lILJDwVfmIOfPpzzR5mCfYOcsG+QE/aNzDR1KnDE\nETb3LlHx9o28PAs2a9a0QjNuBp716wO9ewP9+8e2f8zBn4icDeBnAC0ArAFwGIDJInJOAu2kAFLl\nyF+22WsvC/i3b/e6JURUXigETJxo/xIRUWzSsb5feTk5luo5bpy7KZ8Rt94KDB8e2w37eOb8zQBw\ns6qOiXosH8DLqpqXWFPdxzl/qbN5M7DnnsCWLV63hNJp331tcnKzZl63hChzhULAzJl2d7j8RYEq\nUFpqJcIjX+vWAZ07A/Pm2R3kVFxMEBFlou7dgVNOAXr29Lol7rrlFqBWrbIRwKTW+QvbF0D5deR/\nCD9OWcAPKZ+Ufk2acN4fUSotWQI0bAi0b2/pO3XrAnXq2Id49epWIKBmTXt8jz2ARo2Aww6zOSPF\nxbbO0zvvWFBIRESV82LkLx3uuMOKMq5eXfl+8QR/hQD6lHvs9vDjlAX8kvLJHPz0atw4OPP+2DfI\niZ/7xj33ANu22ffVqwMjRtiHdyhkj5eW2ldREbB1K7BpE7B8OXDUUUCNGhYMvvEGsN9+wO23A5Mm\n2WghxcbPfYO8xb6RedavtxtuLVokdxw/9o399gO6dAFefLHy/eIJ/q4D8HcRWSoiP4rIUgDXhB+n\nLMCRv+wUpOCPKGh+/RX4z39sfaZIIYA2bYBdd7UFfGvUAGSnpJ2y+SPff2+pn9OmAd99Z49feilw\n6KHAww8Dc+em/2ciIvKrKVPKbpxlorvvBl59tfK54DHP+QMAEakBoC2AJgCWAvhRVYuSbKerOOcv\ndUaPBvr1A7791uuWUDo9+qiNODz+uNctIUqvyubhuaVLF0s/uukmS+PMzU3+XKqW1jR0KPDBB3YD\np1s3oGtXm8Objp+LiFKPf8vx698f+PNP4IUXvG5J6nTrZjcSH3ooyXX+RKQVgNWq+kPUY/uJSANV\nneZSe8nH/JL2SenVpAnw3/963Qqi9AqFbA7e3LmpK6gyZoyVHB86FNhlF/fWfBIBjjnGvp55Bhg7\n1s5x1FE2wrh4saWNslAMUXCl4z0qE02eDJx1ltetSK1bbgGOP975+XjSPt8DUH5FjFoAhsTfLAoi\nv6R9+jHPOpMFKe2TfYOcxNs3/vtfu6MevSCvm0pKrDR3//4W+KVK9epAp07Am29a4aazz7a73qn6\nuYKI7xvkxM99o6Agte9RmcqtYi9+7huqlc/7jif4a6aqv+14cP0VQPNYDyAinUVkrojMF5G7K3j+\nDhGZKiJTRGSGiBSLSP3wc7eEH5shIjfH0W5yCUf+slOQgj8it8yfXxaUHXywuwvyAhaM1a8PXHCB\nu8etTO3awPXXAwcdZNstWrj/cxFRenzzDVCvnn2fiveoTLR2LbBypc2JzmR5efblJJ51/mYDuExV\np0Q99jcAQ1W1ypo5IlINwHwAJ8HmC04C0FVVK5yOLiJnAbhVVU8WkVwAwwAcA6AYwEgA15YPRsOv\n45y/FPnHP2zu12OPed0SSqelS4HWrYG//vK6JUTpc9JJwJVXAuPH21IKY8faKJob1q2zwGvkSPvb\nSrdQyEYATzzRisIQUbCsXg0ccoitwTtggI36FRS49x6VqUaPtmvYsWO9bknqhULAbrslv87f8wA+\nF5GbROSM8OjbvwA8F+Pr2wBYoKqLw0VihgM4t5L9u8ECPgA4HFZcZpuqlgAYB6BLHG0nF3DkLzs1\nbAisWWOpJUTZ4M8/bS7ehRcCL79sVTifesq94z/2mAVfXgR+gM0LeuMN4KWXUn9TJxQCJk6svPIc\nEcXnxRcta6BFC+CVV2wt0MjC3pkgVe8bmbq+X0Uqm/8Zc/Cnqm/A1vU7E8DTAE4HcLuqvh7jIZoC\n+CNq+8/wYzsRkToAOgP4JPzQTADHi8geIrIrgDMA7Bdr28kdfgn+/JxnnYlq1AD22isYI3/sG+Qk\nnr4xbJhV4dxlF7uTPmSIVYb78cfk2zF/vi3I7nX13EMOAS6/HHjoodSdIxSyZSs6drTiA34NAPm+\nQU782DdCISvlf3d48lT16sB77wHPPWdrfAbdihVWqTIV7xtuBn9+7BuxqjL4E5GjRSQPAFT1IwCX\nA5gOC9xOFZF6KWjX2QB+UNV14fPOBfAUgG8AfAlgKoCSFJyXKuGXgi+Ufpz3R9nk/feByy4r2953\nX+C112z9vA0bkjt2nz520daoUXLHccODDwL/+peltabCP/9plQhZkILIPQMHAqecYvP8IvbbzwLC\nSy/1702WWEyaZDeMUlWUKptG/ioTy1IPAwA8Cht9A4DXYev8DYKlZj4N4PoYjrMEQLOo7X3Dj1Wk\nK8pSPgEAqvo2gLcBQET6YsdRxB307NkTzZs3BwDUr18frVq1Qn5+PoCySJ3b8W+vXw8sXlyAggLv\n2xPhp99PJm83bpyPpUv90x6n7chjfmkPt/2znZ+fH9P+v/0GrF6dj44dd3y+Sxfg7bcLcOGFwNdf\nJ9ae/v0LMGUK8PHH3v8+AGDatAJ07Qr06ZOPr74Cxo517/irVwOPPFKAevWATZvy0bIlsGaNPz4/\nuM3teLYj/NCebduA556zv9fyz++1VwEOOQS4+eZ8vP22P9ob63ZREfD3vxdgxAhgwIB83HMP8Ndf\nBWjWDMjNded8n31WgFWrgIMOcud4kcf88PuLbBcWFmLdunUAgEWLFsFJlQVfRGQVgKaqui1ceXMl\ngFxVnS8i+wGYoKpVpmCKSHUA82AFX5YB+AlAN1WdU26/3QH8BmBfVd0S9fjeqrpSRJoBGAWgraru\ndA+WBV9Sp1074NlnbV0Zyi69etndst69vW4JUWrdc4+VyK5ojt+mTfZ38MADQPfu8R23uNjW2Xvi\nCeDcyma7p1lREXDEEZYydsYZ7hxT1dJmDzwQuP9+q0I4ZAhw8snuHJ8oW736KjBqFDBiRMXPb9wI\n/O1vNq/4kkvS27ZEzZplKej77GNVkBs3thG6M84AfvnFvYyzUaNsXuS337pzvCAQSbzgSw0A28Pf\ntwWwTFXnA4Cq/gGgfiwNCBdquRHA1wBmARiuqnNEpLeIXBO163kAvooO/MI+EZGZAD4HcH1FgR+l\nFuf8Za+gpH2yb5CTWPpGaakthh6d8hmtbl17/tZbgd92qjVduYED7e/onHPie12q1axpC8H36WOB\noBveeANYtMgC3QYNLL30hRfcOXYq8H2DnPipbxQVAU8/Ddx3n/M+9erZnOWbbgIWL05f2xJRUmID\nCvn5wLXXAl98Ye+RgN1ka9AAmDfPvfO5nfLpp74Rr1iCv1kALgp/3xXA6MgTItIUwPpYT6aqo1T1\nMFU9RFWfDD82KLpojKq+o6qXVvDajqqap6qtVbUg1nOSe/wS/FH6BSX4I0rG99/b2ntHHOG8T+vW\ndvHVvXvsFXDXrLGlcp5/HpCd7sF678wzbV7j67GWb6vE3Lk22jd0qK0rCABXXQX8/DMwbVryxyfK\nVkOH2hqdbdtWvt/RRwN33GE3sUp8Wh1j4UKgUyfg88+tkFavXju/N55/vs1Jdgvn+5WJJe2zA4B/\nA1BYkZUOqjov/NztAI5VVd8MLjPtM3VycoAlS1j0JRt99hkweDDw73973RKi1LnmGiuicNddle9X\nWmopSW3aWFBXlZtvtkDx1VfdaWcqTJ9uRSTmzbMAOBHbttn0gGuusTv50fr3B6ZMsVEJIopPSYml\nT7/yiq1BWpXSUvt7zs+3kXe/ULVriXvvtcJXt93mvDbhTz8BV1wBzJlT8fPx2ndfu8F3wAHuHC8I\nnNI+Y1rkXURyABwKYL6qhqIePwxASFWXutnYZDD4S42SEqBWLUs7qBbLeDFllJ9+Aq6/3u6cEWWi\nrVuBpk2BwkKrnFeV5cttFPCDD6wkuZPZs4ETTrALmL32cq+9qdCrl2V3PPNMYq+/805gwQK7W1/+\nLn4oZHMAJ0ywZSaIKHYff2x/lxMnxp49sGSJzf/77DO7KeO1ZcvsPWbpUpsDnJtb+f6lpUCzZrYw\ne4sWyZ87Lw9Ytcqf2RepksycP6hqSFV/jg78wo/P81PgR6mzYYPlkvsh8AtynnVQNW5sHyR+X6yZ\nfYOcVNU3vvwSOPLI2AI/wIoTDB4M9OgBrF1b8T6qwO23Wxqk3wM/wIpE/POfwK+/xv/a0aNtVO/N\nNyu+uMrJsRtITz+ddDNdx/cNcuKHvqFq82fvuy++wKVpU2DQIEtRXx/zBC1niSy8HnnNO+8ArVpZ\nMPrf/1Yd+AF2vXneee6kfv78s6V8uhn4+aFvJMoHl/IUBBs2cL5fNtt1Vxvp8PtizUSJKr+2XyzO\nOMMuTq65xi7QyvvySyt8csMNrjQx5fbZx4LVqtJey1u1CujZE3j77cqD3JtvBj75xNbwIqLYfPWV\nZV2ddVb8rz3vPODUU5N/DwqFrNJ7x47AsccCM2faTaL58y2rYeZMm9M7ZYqt1TdxIvD115Ydcdxx\nVil8+HBLk69VK/bzujXvj/P9dhRT2meQMO0zNaZPt7tHqVoMmPxt4sSyJT6qVbMUlPPP97ZNRG5Z\nuxZo3tyq48U7323rVpv7d8stwNVXlz2+fbsVjnn+efeWUEiHLVssxWrIkMrTWSNU7b3g4INjSxft\n08emEQwYkHxbibLB8cfbqHm3bom9fvNmKwJz//3x3+Bau9Zu2Lz2mgV2EY0a2U3h6tWdvzZvtqAQ\nsKrC48ZVXaymvKIiuyk1bZrN2UvUWWfZ+3O2XbcklfZJtGEDC71ks7w8S4mrUQPYYw/g73+3eUxv\nvgmE1xMlCqyPP7biCIkUOtllF0t3vPvuHcuSv/KKzXELUuAHAHXqAE8+aSOApaVV7//66xY09+0b\n2/H79AHefRdYuTK5dhJlg++/t/lqF11U9b5Odt3V3qNuuy22JWo2b7a5zOeeazfFRo2y94MjjrAg\n7qijbG7vb7/Zv3Pn2lp906cDU6faKNuPP9r83qOOste0bBlbqmd5NWta4PbZZ/G/NkKVI3/lMfij\nmPhpmYcg51kHVU4O8MMP9kG0cKF9GN12GzByJLD//vbBNGKEjXZ4iX2DnFTWNxJJ+YyWm2vz5bp1\ns4qXK1faHJ3nnkv8mF7q2tUuut57r/L95syx0YRhw8qWdahKkybAxRf7a90/vm+QE6/7Rt++wD33\n2I3XZLRqVfkSNUVFlqZ+2WX2N/rWW0CXLsDvv9vNse7dgfHjbfTu++/tmqAqOTm2bzyvqUiyqZ9L\nllgAmMzIYUW87hvJYPBHMfFT8EfeyMmxlI2cHMvZP+88SwdZtMhGTfr3tzfXG2+0Cd2qiU0QJ0qn\n33+31KTTT0/uONdea8Vi7rrL5gBeeCFw+OHutDHdRCxwvf9+YNOmivfZtg249FK7OI23Et9dd9mi\n924UoQgSvh9mB7f+n3/+2d6bevRwp1233GLXcfffb+1bv96Csuuvt4Dv8cftM37ePJtneMUVO173\nRV8DxCqR15R36qk2crd6dWKvj4z6ZVOVz6pk/Zy/UMj+uPLy4uucibwunedy28CBNpw/aJA356dg\nWLjQRlGGDLF5PRs32ht2bm5yd/6IUuXJJ63fuvHetmiRLWNQXGx9fuLEYPf5rl0tgH344Z2fu+MO\n4JdfKl7WIRaXXWafaffck3w7gyAUsnL78+bx/TCThUKWHrlkSfL/zxdeCHToANx6q3vt++UX+5su\nLbX5+4ccYsFl167+Xv+uSxdLQ73iivhf+8ADNgfx0Ufdb5ffJbXOX5DEE/yFQvZHsGSJpbjss09s\nSxmUllrlw6Ki2F+XyGsir1uxomyBT68+MJ56yi7i/Vimm/xH1dJGrrnG+nCNGtZ3453sTZRKqnah\n9tprVlQhWRMnWpGU4uLECxz4yaJFVihi+nQrGx/xzTfAlVfamoiJLmExcyZw8sk2b2jXXV1prq99\n+SVw5pn2fSb0DarYsGE2Ig4k97k3Z44t0P7bb0Dduu61L/o9Kkify0OGWKZRInP/One2aqdnn+1+\nu/zOKfhLMos42D7+2AI/wC4CXnjBytJWZcoU4JJL4ntdIq+JvO6CC+z72bNtUq0Xf6h+KvhSUFCA\n/Px8r5tBlRCxeT0vvWQXeSLp6T/sG+Skor4xfbrdBDzuOHfOkZdnN+lmz068wIGfNG9uN3Duv9/W\n/wNsWYcrr7TtZNYuzMuzz7LBg4GbbnKhsUlI9fvG2rU2wtmwof3+6tULft/IFvH2jR9+sP/nNWts\n+9NP7QZKzZrxnffJJ21pFDcDPyC471FnnWUB3KZN8f1OIsVejj7a/TYF+Xoja4O/jRttrkKzZla8\nomVLuwsZy6jannvu+McTy+sSeU3kdfXr2wWKl3+o69fbaCVRrCKTvWfNAgoKgHPOAcaO3XEEgchL\n771nhQxiycKIRXSfz83NjLS+e+8FDjvM5h/97W9WLr1bN/sMS9Z991lqW+/e8a39FSQbNtjIw8kn\nA488YheivXpZBcVkKjiS/2zZYlUyCwrsGnOvvSyAO+44e6859NDYjrNwIfDFF7aOntuC+h61xx62\nvuCoUWUDIrFYvNiKUTVpkrq2BVHWpn1efbWlo734YmJ/BKFQ/K9L5DWAffguX25t9eoPtUcPK+px\n+eXenJ+C76mnbBHoggLeSCDvlZRYpdqvv7Yba+Ts9ddteYb27a0QxKRJ7gVrp5xiaXJXXpn8sdI5\nFz8Wmzdb4JebC7z6atncyEmTbCRj8mQrEpQKfqgVkG3eeQf48EPgP/8pe0zV/u8fecQGHHr1qnqO\n7PXX203/J55IaXMD59VXbfmIqqoQR/v4Y0sZ/fzz1LXLzzjnL8qHH1oay5QpwXhTHDrU8pw//NC7\nNpx7rn04n3eed22g4PvHP6wfjxkD7L23162hbPbdd7bm3NSpXrfE/9atAxo3tgXtDzvMghe3PjvH\njLFKqbNnW1GGRK1ebeXsly6195Yrr7Q5TSUlzl9btgD//rdltuTlWcqeWz/X1q02x6hJE7vpVX50\nuV8/C6S//Ta5n7sioZCN0i5caD8Xi8ukR7t2Nppd0dyy2bMty6BZM1sf1+nzb9kyu1kwd66lj1KZ\nJUtsjvby5bHffLrnHksTffDB1LbNr7jIe9jixVaKfujQ4LwZHnbYjosHe8FPSz0EeW2VbPfgg3Yj\n4ZRTyuZEuIl9g5yU7xvvvZfc2n7ZZM6csrXBfvvNMljckp8PNGhgc6MSNWeOjUouWWIZPStXWqG0\nmjXtwq9BA8s2aNYMOPhgG+lt1coC2vXrgdLSAsyYYYGYG7Zvt3TWBg1sTmNFacV33WX/9u/vzjmj\nffCBVXUsKSmrFUCJifUzpbDQ+t8ZZ1T8fMuWtgRSixa28PnIkRXv99xz9r7EwG9nTZta6mw8H/Op\nXNw9yNcbWTXnr6TE/qj69AGOOcbr1sTu0EOBBQvKSvN6wU8FXyi4RGwtoW3bgNNOA0aP9s9NBcoe\nW7ZYNsXjj3vdkmBIZZEIERsteeghC5jiWTYiOqXuwQct0Jozx9o4YEDVN3hDIZtDNHOmjcT8/e82\nV6t798TXBCsuttdXr243GJwW565e3dLRjj4aOOkk965Jpk2zqSLNm1u11gMPDE5RjyAbNMhSOisb\nxa1d26Y/dO5sSxace65VUK9Tx55fs8b68LRp6WlzEEUWfD/11Kr3VbW5yqko9hJ0WZX2+dhjdsfg\nm2+8C6IS1aSJ3TVq1syb8x90kKWoHHywN+enzKJqC85OmmRzroIyCk+Z4aOP7GJt9GivWxIcic5Z\nj4WqjcT16+c8clLesmXAVVfZBfOQIXaTNNm5+L/8YjeIc3Nt+Y8994zv5ygttYv6FSuAESPsYr8q\nH35o65BNmWJVQJMxdy7QqZMFvqefDlx3HdCoEfDss8kdlyoXCtn84ZkzYy8ssnatze2bNs3Wxm3d\n2m5i/PGHBYBUsfnzLVvgzz+rvo7/5Re7sbJ4cVqa5ktZn/Y5YQLw8ss2aT1ogR9gqZ/z53t3fj+l\nfVLwidhyJ0ceaWtfbdrkdYvSIxSydZZCIa9bkt2Y8hm/nBxbmiEVN2pEbLSqb18LBKvyr3/ZxXKb\nNjZPL1JFMZE2Rr+mdWsbKdh3X0vNGzUq9uOo2tzFP/6w9sUS+AG2JM5xxyW/kPfChTYa8sQTdsyc\nHODOO21ttNLS5I5dlWx/Xxs6FDjxxPgqSu6xh73u/vvt/+2hhyxov/HG1LUzExx6qP3ufvyx6n1T\nmfIZdAEMg+K3fr2lYQwaFNwy84ce6t28P1V/pX0GOc+ayojY3fUDD7RlILZsSf6Yfu4bM2bYvKMO\nHewic8MGr1uUXSJ9Y/VqywDp0sXT5lA5F11kI2bff++8Tyhko3133mlpu48+Gv/6aRWJft/YZReb\nd/Xuu7YExfXXV31zStWCtxkzrIBMvIvWv/ii9clPPom76QCsyM3JJ9s8wp49yx4/8kgbTZwwIbHj\nxiIUsiIcHTrYvMtMCwCr+kxRtc+xa6+N/9gidm06Zoz1uQ0brFBRpv0O3RZJ/axKqoM/P19vVCXj\ngz9VS33o3DnYlSq9LPqydauNlsZ6J5MoVtWqWYpLo0Z2Mb5tm9ctSo0lS+w9aOtWuwv/66828f/B\nBy1Vi9Lno4/s/8IvN7PIVK8O3H23c3n7CRMsNbR6dSuu0bZtatvTqZOl5IVCdrPGaaRB1eYs/vCD\nFfFIZGQ0J8dS/66/3tLZ4rFypQV+vXrtPGoUCS7iKY0frxkzLK2utNTSHlN5Lj/66SebJ3rSSYkf\nIxSyzz5VFuiJRST4qypLgCN/lVDVjPqyH6nMu++qtmypummTBtoXX6iedpo3516+XHXvvb05N2WH\noiLVCy5QPecc1W3bvG6Nu1asUG3RQvWRR1SPOkq1Zk3VI49U/eEH1T59VBs3Vj36aNXnn1ddtszr\n1ma+445THTHC61ZQRbZuVW3aVHXy5LLHtm9XfeAB1UaNVD/7zJt2ffSRasOGqg89ZO2J9thjqnl5\nqqtWJX+exx9XPfFE1ZKS2PZfu1a1VSvV++5z3mfRItU990zd++q336rWqmXvawccYP9P99238+8p\nU/Xsqfr008kdY8OGss+Go46ybXJWWqrarJnqjBnO+5SUqObkqK5enb52+VE4JtopVsrokb9ffwVu\nvwtcOY4AACAASURBVB0YNiz+NAy/8TLtk/P9KNVq1LD5D6q24PPatZkxh2TdOpvP0aUL8PDDltI2\nbpyNEhx3HPDMMzZH6MknbTTj8MOtCuqQIXY3GeB8GjctXGjvo6ed5nVLqCK1awN33GGFXwD7v2rf\n3oqhFBZadUQvXHihnf+nn6w9c+fa3+NNN9nC3t98E39xmIrcc49VC33mmar33bjRiuN07Fh51dr9\n97fqp/HMX4zHkCFWsGbcOBspnTbNflft23tbp6C8VLyPrl1rI1DRqbaJyMkp+2zgmoxVE7FMvspS\nP+fPtwq+DRqkr12BUlFEGOQvhEf+tm9XbdNG9YUXXAmePVdUpFq7tuqWLek/908/2ciEX4wZM8br\nJlCKbN2qevLJqrvvrlqjRvx3Qf3UN0Ih1XbtVG++2e5UxmLTJtXhw1XPOst+BxddZHfTE/ld0I7G\njBmjjz+uet11XreEKrNxo+pee9mISoMGqq++GvvfT6Jifd8oLVV97TVrV/36qoDq4Ye7+3e5aJFl\n2kSPfpa3ZYtqp06qV10V2yjhwIGqF1/sXhsj1qyx96kVK3Z8vLRU9ZVX7P/x9ddT//9XlQ0bLNvC\n7c+UAQNUL700+fZR/MaMUW3d2vn5IUNS0+d3bMOY1J7ABci2kb9HHrE7cTfd5HVL3FGjhq3b88sv\n6T+3n4q9UGarXduqn4VCdgd8+nS7Cx60Ua+tW22U4vDDgeefj33NsF13BS65xIpGLFgA7LefjVYV\nF3MuSLJUWeUzCEpLrYjLP/9pd+4vuyzxNffcJmKFPV57zTJiAPtMdvPvcv/9rQDMpZdWXGgmsoD8\n3nsDr78eW/Xyiy6ykT+3i0y9845Va9577x0fF7H5i2PH2jqMXboAq1a5e+54DB9unyVuvo+qAgMH\nJlbohZLXoYNlzSxaVPHznO9XuYwM/goKgLfftg8Pv3xouMGroi9+S/vMz8/3ugmUQkcfbdXjata0\n4Oenn6z0eteuFhRt3+78Wj/0jaIiu9iK5+KsInvvbTexjjzSjlGrllVGpcTsvns+tm0D2rXzuiVU\nmZkzrYgJAPz2W3pueMT7vnH66fZ3WbOm+4veA/Ze17YtcNttOz5eUgL06GHvB0OGVL6geLQGDWwp\ngk8/da+NsQQ/LVva+sQHH2zFer7+2r3zx2rqVCvIEwlQDzssvv8vp74xbpz9P3TokHwbKX41agBn\nn21VfyuSjuDPD9cbicrI4O/yy4G33gIaNvS6Je7yaq2/9es58kfpEz3/YeZMq6D366+2sOvTT9ta\nStdea/ukev2qeEUuzkTiuzhzkpNj8wPHjbM77FdckbkVUVPtvfes8mEm3RDMRHl5dnGeqsDKDemY\no/XSS8C335bNayottYqeq1fbwvDxLnHRvbtVFHXL2LF2AV5V8FO7NtC/v40SXn21BbRbt7rXjsrM\nnm3zIgcNss+QCy8EDjnElr9IViTw5fuJd5yWfCgutnmnf/tb+tsUFBkZ/J1zjpXyzjReFX3ZsMFf\nI39BXluFYlN+sea99ioL+CZPttSo664DDjjAiiTMmGGpoa+8UuBZimjk4mzVqsQuzpzk5FhxmPff\nB+rWBbp1s9HFVMjU4jLr1gFvvlkQ6OV+soUXxS8S+UxJ5aL3gN1wff99e9+bN8/SwWfPBj7/3NYi\njNdZZ9kC9kuXutO+yNp2sQY/J51kxWD+/BNo08Zu7KXy/WbBAiu29cwzlnaak2NrNy5YYIForCrq\nGytWWBptjx7utZfid8opFuRFMgUi5s61Nb1Tfd0a5GvRjAz+xo3LvIsXwNuRPz8Ff5TdmjcH7r23\nbEFlVbvZ07ChzfHt0CH9f/+RRZ7nzbM0lEQuzqoSqYi6bZuNAJaUuHv8UMguZjt29OZ3mCqhkKWc\nhUI28pApP1cmS3VgFRRt29oNpbw84OOPbQ5gotkOderYPOThw5Nv119/WQpnvMFPgwZ2Y+z224ET\nTrC1Tjt2BI4/3t2/y8WLbe3Dhx+2Ec+IOnWs+vuddyZXP+Htty2grF8/+bZS4nbZxQL8ESN2fJzz\n/aqWkcHf3LmZWRjBy5E/P6V9BjnPmtwjYvNunnrKLiiKigDVfMyYAYwfn962PPCApWf+5z/upBQ5\nqVUL+OQTu/Pcq5e7aa8DBtjIQqTQzs032935ICspAe66yy4GgXwWzaEK+fkz5bTTyv7O581Lrv9e\ndpk7qZ9vvWUplIncFBaxpREGDrRRyOJi+5nc+rtcutRGGfv0sffI8vLyLCjs1q3y+eMR5ftGaaml\nkbLQiz9UlPqZruDPz+8bVcnI4M+v8wSS1bChXcysXp3e83Lkj/zuyCPtQ71mTUsRvfLK9BUX6NfP\nPny++io9d4J32cVSvxYssJFOW+EmcWvWWGXBIUNsPkzNmvbvbrvZfJm8POCxx/y1ZlcsFi8GOnWy\nAPbww/09h4zISatWZQWwku2/+fnA8uV2gzxRJSXuBD+dO5cVs1K1v9Nk38tWrrQRv6uvtptXTm64\nAWjUyILAeH3zDbDHHhxZ8oszz9w522/yZCscR84yMvjL1EUyRbyp+Om3gi9BzrOm1IjMExowoAC/\n/mrFPdJRXOCll4DBg4HRo3cud55KdevaKOOkSZbClOhF09dfA0cdZW2fNs3mBI0bZ/++8IKV0R40\nyC6qOna0D9T+/SMjaf41bBhwzDF2YVBQAPz4o/WNTP1soOT4+TPFzTmQ1atbJdFkRv+++sreL5K9\nuI4Usxo/3v5Gn38euPjixG9ur11rKYBduti0gMqI2Ojlu+8CY8ZUvm/5vsFCL/6y++5A+/ZWGA6w\nDKAZM4DWrVN/bj+/b1QlI4O/TP5w9yL1028FX4gqkpNjd8ZzcsqKC/zxhxUXmDHD3XOFQrYe4VNP\nWeDXpIm7x4/FbrtZ0YHRo21JiHhs2mR3v3v1siVxXnjB5sOUn2tVrZoVm3nxRWDJEgv8FiywC7/2\n7e3xZcv8Uyhm/Xqbh/Too/a7uesuu+CN7htEQePmHMhI1c9EbxgNHGjFttwQ+bk6dLAbTs2a2Wjg\nf/4T33FCIVt+Iz/fshRi0bChBYCXXx57wPnnn1bltFu3+NpHqRWd+jlrlhWE43t95USTHWf3GRHR\nTPuZoj3+uF249euXvnO2awc8+6xd7BEFiaoFN3fdZfPybrop8XX3Iv77X6sovHKljcRPmuTtB82K\nFXbRc/nlVvm0Kv/9r+3btq0Fb4mkqhYVWdA5fLiloBYXA1u2WHrlxIne/D5++MECv86d7f1q113T\n3wYiv1O11NE334z/M/33321E5fffLfsgFQoKbE7gaafZ33FVc6g3b7bAr0ULC0zjHZG7/XbLZPj4\n46pf++ij9n77yivxnYNSa/ly++xZvtymL4wbZ6O6BIgIVHWnnp3WkT8R6Swic0VkvojcXcHzd4jI\nVBGZIiIzRKRYROqHn7tNRGaKyHQReV9EaqWz7X7h1cifn9I+iWIlYvP/Jk60VMAzzrCRqnj9+ivQ\nt6/Nfzv33LI7xelahLoyDRtaIDZ4sAVzTrZvBx58EDjvPLt59O67ic9RrFnTLrjeeceqm27daoUQ\nZs2y3/Hw4XaTKh2Kiuznuugi+/lfe42BH5ETkcTX/HvjDXttqgI/wG5kTZ9uf9dHHVV58a5t22zU\nZ//97e8+kVTMfv3s/f3NNyvfr7jYfv7eveM/B6XWPvtYZsd337HSZ6zSFvyJSDUALwM4DUAugG4i\n0iJ6H1V9RlVbq+rfANwLoEBV14lIEwA3Afibqh4JoAaArulqu594sdyD3wq+BDnPmlLLqW8cfLDN\nl2nTxu5cf/551cdassTmobRpY3fIly2zO8vz57tXgMEtTZpYAPjcc8Drr+/8/KxZNtJXWGhfF1zg\n3rmPPrqs2E5enlUU/Oc/bZ2l7t0thStV6xL+8ouljE2eDEydCpx9tvO+fN8gJ9nWNy69tKxCcqyK\niuwGUzqqXO62m6VkPvecVRW9914L9Mq35+KLy/ZNNKOjdm27MXjffRUXwon0jf/8pywtlfwnkvqZ\nzuAvyO8b6Rz5awNggaouVtUiAMMBnFvJ/t0ADIvarg6grojUALArAJeWKg2Wgw+2u1Rur/FVGb8F\nf0SJqFkT+Mc/bKmEW2+1i5hNm3acr7Z6tRU4yc+3AG/6dEu1XrIEePllCzR23z39i1DHYv/9LQB8\n7DFLfQmF7K75E0/Yz3P99bYe0j77uHve6IIUEybYnfFRoyxIbt/ezt+kif2+x41zZ3kKVbvga9fO\nUj2//NL9n4soUx1wgN1I/uqr2F/z+edWBbhly9S1q7xzz7W523Pm2E246dPt8ZIS+7svKbERzBo1\nkjvP4Yfb+/yll+4cZEZECr2QP51/vmWhzJ5tFXKpcmmb8yciFwA4TVWvCW9fBqCNqu5UkFdE6gD4\nE8BBqrou/NjNAPoC2Azga1WtcHnRTJ/zB9jdp4IC4MADU3+ukhJbW6yoKPm5UkR+sX69zf+bONEC\niUWLylIFTz/9/9u792irqvvQ498fj2hUBN8ZFRFDihH0io/rMzYn6q1Y44M4EoU8IPeWmHvbiGli\n7XPYjNLERm99tderiVFrqiHW1MetQ63VE6XDtxwFRMTgsRCfCfKKBkXm/WOtE45wFnI4e++z1trf\nzxgM9lx77bXm1t/ZnN+evzlnNqF/8uTmbNbebIsWwac+lS108vLL2ft6+OHB/ca6uzsrBb3ppmxV\nvqlTs3mTQ4ZkSXZ/Euju7qyU9403YM6ccoy8SlVz1VXZlzE33/zB58LGLRQGY7GTlLIS8/PPz7Zw\nePTR7Mute+5p3Gd0StlKoePGwSWXvP+5pUuz5HPZsmxhLJXTxInZFISurvJ8KTvYSjHnrx9OBeb2\nSvxGkY0S7gv8FrBTREwbxP4NqlaWfq5Zk024NvFTnYwcmc15++IXN46kv/VW9s3hnDnZvLgqJn6Q\nfYv93e9miR9kX9y89dbg9mns2GwxmmeeyZbk3rAh24Pv2GOz/xc77piVb40cmc1D3GUX2HVX2G23\nbN/GPfbI5jbusUf2y9mDD2afSWPGDO77kqrqs5/Nfha3ZoXe55/PVkz+zGea36++9GwM/8AD2QrL\n//qv2f6kjSwnj8jm/c2Zk+3l19v3vpctkmXiV149lTvd3XDccYO/8nTZDXCwvF9+DvT+p3p0fqwv\nZ/P+ks8TgaUppRUAEfET4Bjgpr5ePGPGDMaOHQvAqFGjmDRpEh0dHcDGGt0qt3fcERYv7mDy5Obf\n7557OvnQhwDK8/67uro477zzStMf2+VpX3bZZf36eT/88E7GjYP//M8OJkyAX/+6k87O8ryfbW1P\nmdLB3/0dLFjQyZgxMHFiufp35pkdXH45QCdDhsAdd3Rw+OHw0EOdpATHHttBSjB3btY+5piOvNSz\nk7/4C9iwoYPnnoMbb+xkwoStu3/P4zK8f9vlavccK0t/WtHefXc44IBOvvMd+Pa3t3z+nXd25Atn\nDW7/f/rTTt5+G6CDxYv79/O/Ne358zs57zyYMaODri5YuLCTJ57o4gc/OI8HHyzX/z/b728vWABv\nvNHJhg3w7LMdLFyY/XvezPv39/eNVrS7urpYuXIlAN3d3RRpZdnnUGAxcALwCvAYMDWltGiT80YC\nS4HRKaW382NHANcC/xVYB1wHPJ5S2mzB3XYo+7ziimzFz1YsNzx/flbmsWBB8++1tTo7O38T7FJv\n2xIba9ZkC6JMnFivUpEyv681a7JvZ599NptDtLVzJ7f1deDnhoq1a2zMmZPNnd3S3L+3385G2B99\ntDVTTbZkID///XHBBVn5/O23w4UXdjJ3bgf339/4+6hxWhUbvVXhc6Oo7LOl+/xFxGTgcrJy02tT\nShdFxDlASildk58znWxu4LRNXnsh2Yjgu8A84PfzhWM2vUftk7+7785q0u+7r/n3mjs3+yDc0nLL\nktRf25qcljmplarkrbeyVXkXLSpeMOnGG7NFVe6+u7V9K9KKn/933skWq5o2DX74Q5g1C6ZPb869\n1Dj+27C5UiR/rdAOyd+LL8InP5lttNpsd90FV16ZzQ2QJEn1MX06HHpoluD05Zhjsi+AT9/S2uw1\n9NRT2SIv772XLUr1H/9hQqHqqdqCL9qCMWOyle5asYlyGbd56KlzljZlbKiIsaEi7RwbW9rw/emn\nsy+ZTzmltX0qg41bPnTy3HPZiJLUW5U/N0z+Kmjo0GzFuyVLmn+v1auzVfgkSVK9HH98toVBXyuI\nX301zJw58H30qujAA7M/Q4dmc8jcUkZ1YtlnRZ15Jpx1Fnzuc829z3e/m40yXnxxc+8jSZJa7+tf\nz77k/da3Nh5bswb23Tdb9G3vvQevb4PJOWSqOss+a2b8+GzFz2YrY9mnJElqjJ7Sz97fm990E3R0\ntG/iB1nCd9RRJn6qH5O/imrVRu9lLPuscp21msvYUBFjQ0XaPTYOOywrb3zssaydElx1FXz1q4Pb\nrzJo99hQsSrHhslfRe2/vyN/kiRpYCLgC1/ItjWAbE+/tWvhxBMHt1+SmsM5fxX1y19mG66uXJl9\ncDfLGWdkS0FPmdK8e0iSpMHzs59l2zosX54t8jJxIpx//mD3StJAOOevZnbbDYYPh9dfb+59HPmT\nJKnexo3LvlC+5Ra47TaYMWOweySpWUz+KqwVi76UMfmrcp21msvYUBFjQ0WMjcznP5/N8zvySNh+\n+8HuTTkYGypS5dgw+auwViz6UsYFXyRJUmOdckq2vcH998Nxx2WPJdWPc/4q7KKLsrl/zdyDb489\nsn1u9tyzefeQJEmD6+GH4Xd+B9avz6aVPPhgttWBpGpyzl8NjR/f3JG/lLKyT0f+JEmqtwMPzBZ6\nGT4cJkzIHkuqH5O/Cmv2dg/r1mUriZat9r/KddZqLmNDRYwNFTE2MiNGwEMPZSN+Dz3k5uZgbKhY\nlWNj2GB3QNtu3Djo7oZ3382+qWu0Mi72IkmSmmPECEs9pbpzzl/FffSjcM898Nu/3fhrL1kCJ58M\nL7zQ+GtLkiRJag7n/NVUM0s/HfmTJEmS6sPkr+KauehLWZO/KtdZq7mMDRUxNlTE2FARY0NFqhwb\nJn8V18yRP/f4kyRJkurDOX8Vd999MHs2NOMLiOuvhwcegBtuaPy1JUmSJDWHc/5qav/9m1f26cif\nJEmSVB8mfxW3997Z3LzVqxt/bef8qWqMDRUxNlTE2FARY0NFqhwbJn8VN2RIts3DkiWNv3ZZkz9J\nkiRJ/eecvxo46yw4/XSYNq2x1/3KV+Cww+Cccxp7XUmSJEnN45y/GmvWdg+O/EmSJEn1YfJXA83a\n7qGsC75Uuc5azWVsqIixoSLGhooYGypS5dgw+auB8eObk/w58idJkiTVh3P+amDVqmzVzzVrIDar\n7N12EyfCnDlw4IGNu6YkSZKk5nLOX42NHAk77QQvv9zY65a17FOSJElS/5n81UQzFn0pa9lnleus\n1VzGhooYGypibKiIsaEiVY4Nk7+aaPSiLxs2wK9+lY0oSpIkSao+5/zVxMUXZ2Wfl17amOutWgX7\n7JOVfkqSJEmqDuf81dz++ze27LOsJZ+SJEmSto3JX000uuyzzIu9VLnOWs1lbKiIsaEixoaKGBsq\nUuXYaGnyFxGTI+K5iHg+Ii7o4/lvRsS8iHgqIuZHxPqIGBUR43sdnxcRqyLi3Fb2vez22w+WL4d3\n3mnM9Rz5kyRJkuqlZXP+ImII8DxwAvAy8DhwdkrpuYLzPw2cl1I6sY/rLAeOTCkt6+N1bTnnD7IV\nP2+/HQ44YODXuusuuOIKuPvugV9LkiRJUuuUYc7fEcCSlNJLKaV3gR8Bp2/h/KnAzX0cPxH4WV+J\nX7tr5HYPq1c78idJkiTVSSuTv72B3gnb8vzYZiLiw8Bk4NY+nj6LvpPCttfIeX9lLvuscp21msvY\nUBFjQ0WMDRUxNlSkyrFR1gVfTgXmppRW9j4YEcOB04BbBqVXJdfI5K/MC75IkiRJ6r9hLbzXz4Ex\nvdqj82N9OZu+R/dOBp5MKb2xpRvNmDGDsWPHAjBq1CgmTZpER0cHsDFTr2N7/Hi48spOOjsHfr1V\nqzoYObJc7693u0dZ+mO7HO2eY2Xpj+3ytDs6OkrVH9u2bVej3aMs/bFdjnbPsbL0p7Ozk66uLlau\nzMbNuru7KdLKBV+GAovJFnx5BXgMmJpSWrTJeSOBpcDolNLbmzx3M3B3SumGLdynbRd8eeUVOPhg\neP31gV/ra1+Dj30MZs0a+LUkSZIktc6gL/iSUnoP+EPgXmAh8KOU0qKIOCcivtLr1DOAe/pI/HYg\nW+zlJ63qc9V85CPw61/Dm28O/FplXvBl02/jpB7GhooYGypibKiIsaEiVY6NVpZ9klK6G9h/k2NX\nb9K+AdhsZC+l9BawR1M7WHERG1f8PPLIgV2rzAu+SJIkSeq/lpV9tko7l30CfP7zcNJJ8KUvDew6\nxx8Pf/7ncMIJjemXJEmSpNYY9LJPtUaj9vpz5E+SJEmqF5O/mmnUdg9lTv6qXGet5jI2VMTYUBFj\nQ0WMDRWpcmyY/NVMo0b+3OdPkiRJqhfn/NXM2rWw557Z30MGkNpvt102+rf99o3rmyRJkqTmc85f\nm9hpJ9hlF1i+fNuvsW4dpJQlgJIkSZLqweSvhgY6769nvl9s9l1BOVS5zlrNZWyoiLGhIsaGihgb\nKlLl2DD5q6FGJX+SJEmS6sM5fzV06aXw4otwxRXb9vonn4SZM+GppxrbL0mSJEnN55y/NuLInyRJ\nkqRNmfzV0EC3eyj7Ng9VrrNWcxkbKmJsqIixoSLGhopUOTZM/mpo7Fh49VV4++1te70jf5IkSVL9\nOOevpiZMgDlz4KCD+v/aK66AJUvgyisb3y9JkiRJzeWcvzaz335w552wZk3/X1v2sk9JkiRJ/Wfy\nV0Nr1sDjj8Nf/iUcd1z/E8Cyl31Wuc5azWVsqIixoSLGhooYGypS5dgw+auhBQtgxQrYsAGefRYW\nLuzf6x35kyRJkurHOX81tGYNfOITMH8+7L13lgCOGLH1rz/7bDj9dJg6tXl9lCRJktQczvlrIyNG\nwNy5cN11sG5dNgLYH2Uv+5QkSZLUfyZ/NTViBEyfDqedBrNn9++1ZS/7rHKdtZrL2FARY0NFjA0V\nMTZUpMqxYfJXc7NnZyOAS5Zs/Wsc+ZMkSZLqxzl/beBv/xYefhhuu23rzt9nn6xsdN99m9svSZIk\nSY3nnL82NmsWPPMM/Pu/b935q1c78idJkiTVjclfG9h+e7jkEvj612H9+i2fu2EDrF3bv9VBW63K\nddZqLmNDRYwNFTE2VMTYUJEqx4bJX5uYMgV23RWuvXbL561dCzvsAEOHtqZfkiRJklrDOX9tZN48\nOPlkWLy4uKxz2TI4+mhYvry1fZMkSZLUGM75E4ccAqeeCn/918XnuNKnJEmSVE8mf21m9my4/vri\nrR/KvscfVLvOWs1lbKiIsaEixoaKGBsqUuXYMPlrM3vtBeefD9/8Zt/PO/InSZIk1ZNz/trQunUw\nYQJcfTWceOL7n5szB269FX7848HpmyRJkqSBcc6ffmO77eDii/ve+sGRP0mSJKmeTP7a1JQpsNtu\n8P3vv/94FZK/KtdZq7mMDRUxNlTE2FARY0NFqhwbJn9tKgIuuwz+6q9g5cqNx6uw4IskSZKk/nPO\nX5ubOTMb6bvkkqx97rkwbhzMmjW4/ZIkSZK0bUox5y8iJkfEcxHxfERc0Mfz34yIeRHxVETMj4j1\nETEqf25kRNwSEYsiYmFEHNnKvtfVpls/OPInSZIk1VPLkr+IGAL8PXASMBGYGhEf731OSumSlNIh\nKaVDgT8FOlNKPUWJlwN3pZQOAA4GFrWq73W2117wx3+8cesH5/ypyowNFTE2VMTYUBFjQ0WqHBut\nHPk7AliSUnoppfQu8CPg9C2cPxW4GSAidgaOSyldB5BSWp9SWt3sDreLWbNgwQK4775qJH9dXV2D\n3QWVlLGhIsaGihgbKmJsqEiVY6OVyd/ewLJe7eX5sc1ExIeBycCt+aH9gF9ExHV5Seg1+TlqgN5b\nP6xYUf6yz5W9V6iRejE2VMTYUBFjQ0WMDRWpcmyUdbXPU4G5vUo+hwGHAv+Ql4S+BfzJYHWujqZM\ngd13h6efhqVLYc2awe6RJEmSpEYa1sJ7/RwY06s9Oj/Wl7PJSz5zy4FlKaUn8vY/A5stGNNjxowZ\njB07FoBRo0YxadIkOjo6gI01urY3b8+eDZ/4RCfTpsFBB3Xw0EPw5JPl6V9P+5FHHqFHGfpjuzzt\nRx55hM7OztL0x3Z52t3d3aXqj+3ytLu7u0vVH9vlafv7hu2idhl/3+jq6vrNiGTP51pfWrbVQ0QM\nBRYDJwCvAI8BU1NKizY5bySwFBidUnq71/GfAjNTSs9HxIXADimlvlYMdZ8HSZIkSW2tr60eWjby\nl1J6LyL+ELiXrNz02pTSoog4J3s6XZOfegZwT+/EL3cu8E8RMZwsOfxywX02e5OSJEmS1O5qt8m7\nJEmSJGlzQwa7A5IkSZKk5jP5U6lFxLUR8VpEPNPr2C4RcW9ELI6Ie/J5omojETE6Iu6PiIURMT8i\nzs2PGxttLiK2i4hHI2JeHhsX5seNDQEQEUPybaPuyNvGhgCIiO6IeDr//HgsP2Z8tLmIGBkRt0TE\novz3jiOrHBcmfyq764CTNjn2J8B9KaX9gfuBP215rzTY1gN/lFKaCBwN/EFEfBxjo+2llNYBn0op\nHQJMAk6OiCMwNrTRLODZXm1jQz02AB0ppUNSSkfkx4wPXQ7clVI6ADgYeI4Kx4XJn0otpTQXeHOT\nw6cDN+SPbyBbJEhtJKX0akqpK3+8FlhEtn2MsSFSSm/lD7cjW9gsYWyIrGoA+D3g+70OGxvqEWz+\nu7Hx0cYiYmfguJTSdQAppfUppVVUOC5M/lRFe6aUXoMsCQD2HOT+aBBFxFiyEZ5HgL2MDeVldBjf\n0wAABZZJREFUffOAV4F/Syk9jrGhzKXA+WRfCPQwNtQjAf8WEY9HxO/nx4yP9rYf8IuIuC4vF78m\nInagwnFh8qc6cMnaNhUROwH/DMzKRwA3jQVjow2llDbkZZ+jgSMiYiLGRtuLiFOA1/KqgS1tC2Vs\ntK9jU0qHko0O/0FEHIefHe1uGHAo8A95bPyKrOSzsnFh8qcqei0i9gKIiI8Arw9yfzQIImIYWeJ3\nY0rp9vywsaHfSCmtBjqByRgbgmOB0yJiKXAzcHxE3Ai8amwIIKX0Sv73G8BtwBH42dHulgPLUkpP\n5O1byZLBysaFyZ+qIHj/t7R3ADPyx9OB2zd9gdrCD4BnU0qX9zpmbLS5iNi9Z9W1iPgw8N/I5oQa\nG20upfRnKaUxKaWPAmcD96eUvgjcibHR9iJih7yahIjYEfhdYD5+drS1vLRzWUSMzw+dACykwnHh\nJu8qtYi4CegAdgNeAy4k+zbuFmAf4CXgcymllYPVR7VeRBwLPEj2D3PK//wZ8BjwY4yNthURB5FN\nvh+S/5mTUvqbiNgVY0O5iPgk8I2U0mnGhgAiYj/gX8j+PRkG/FNK6SLjQxFxMNkiUcOBpcCXgaFU\nNC5M/iRJkiSpDVj2KUmSJEltwORPkiRJktqAyZ8kSZIktQGTP0mSJElqAyZ/kiRJktQGTP4kSZIk\nqQ2Y/EmSJElSGzD5kyTVVkQ8GhEfi4j9IuLJDzh3Q0SsiYjV+d8rGnD/6RHx0ECvI0lSI5j8SZJq\nKSKGAWNSSi8AhwFbTP6ABPyXlNLOKaURKaVdG9GN/Lrb9uKIoQ3ogyRJgMmfJKm+DgKezR8fDsz7\ngPMj/7P5ExGfjoh5EfFmRMyNiIN6PXdBRLyQjxguiIgz8uMfB64Cju49khgRD0TEf+/1+veNDuYj\nkP8rIp4Hnu+5VkTcGxG/jIhFEfHZXuf/XkQszO+/LCL+aOv/E0mS2onJnySpViJiRkS8CcwlS7xW\nAN8ALoqIFRGxbz+vdwhwLTAT2BW4GrgjIobnp7wAHJtS2hn4FvDDiNgrpfQc8FXg4a0YSdx0dPB0\n4AhgQkTsANwL/BDYHTgb+D95cgnwfWBmfv8Dgfv78/4kSe3D5E+SVCsppetTSruQlXkeBRwMzE8p\njUwp7ZpSemkLL38qH91bERGX5cdmAv83pfREytwIrMuvTUrp1pTSa/njW4AlZInbQHw7pbQypbQO\n+DTwYkrpH/P7Pw3cCvSM/r0DTIyIESmlVSmlrgHeW5JUUyZ/kqTaiIhd8uRtJXA00AksBvbPE7pz\nP+ASh6SUdsmTxPPyY/sC38hfvyIfVRwN/FZ+zy/1Kgl9E5hINkI3EMt7Pd4XOGqT+08D9sqfPxM4\nBXgpLyk9aoD3liTV1LDB7oAkSY2SUnoT2CUizgI6Ukr/MyJ+AlyZUnpgKy7R15y/ZcDfpJS+s9nJ\nEWOAa4BPpZQezo/N63WdvhZ7+RWwQ6/2R/p6K5vcvzOldFJfHU4pPQmckS8O8zXgx8CYvs6VJLU3\nR/4kSXV0GPBU/viQXo+3xfeAr0bEEQARsWO+yMqOwI7ABuAXETEkIr5MNu+ux2vA6F7zAwG6gM9E\nxIcj4mPA//iA+/8/YHxEfCEihkXE8Ig4PF8EZnhETIuInVNK7wFrgPcG8F4lSTVm8idJqqNDyebv\n7QqsTymt2orX9LklQz6yNhP4+3zxmOeB6flzi4D/DTwCvEpW8jm318vvBxYCr0bE6/mxS4F38/Ov\nI1vIpbAfKaW1wO+SLfTycv7nIuBD+SlfBF7MS12/QlYSKknSZiKlbd5+SJIkSZJUEY78SZIkSVIb\nMPmTJEmSpDZg8idJkiRJbcDkT5IkSZLagMmfJEmSJLUBkz9JkiRJagMmf5IkSZLUBkz+JEmSJKkN\nmPxJkiRJUhv4/26hKlXHpAluAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8fda024ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "plt.plot(range(1,len(scores)+1), scores, '.-')\n",
    "plt.axis(\"tight\")\n",
    "plt.title('Feature Selection', fontsize=14)\n",
    "plt.xlabel('# Features', fontsize=12)\n",
    "plt.ylabel('Score', fontsize=12)\n",
    "plt.grid();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Does the cross validation score not depend on WHICH features are selected? We put the parameters in order of most highly correlated to lowest. \n",
    "\n",
    "Based on the plot, a feature space of 10 dimensions (the top 10 most highly correlating features) provides the most reliable result while avoiding overfit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best features:\t ['Title_Mr' 'Sex_male' 'Sex_female' 'Title_Mrs' 'Title_Miss' 'Pclass_3'\n",
      " 'Pclass_1' 'Fare' 'Embarked_C' 'Embarked_S']\n"
     ]
    }
   ],
   "source": [
    "best_features=param_df.index[1:10+1].values # ignore the first column which is survives\n",
    "print('Best features:\\t',best_features)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking at them in histograms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9d28d30>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fda237358>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fda076ef0>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9cee4a8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9e0b5f8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9ed1588>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9e51320>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9e30ef0>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9cd5e80>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9ca42e8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9d52860>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f8fd9f974a8>]], dtype=object)"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIsAAANwCAYAAACvfeo6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nPCXSMP8AbAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8fd9e24048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df[best_features].hist(figsize=(20,15))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating the train and test datasets within the original training data to experiment with different parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X = df[best_features]\n",
    "y = df['Survived']\n",
    "\n",
    "# Add X to y for a dataframe ready for the final model training. \n",
    "\n",
    "train_final = pd.concat([X,y], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Title_Mr', 'Sex_male', 'Sex_female', 'Title_Mrs', 'Title_Miss', 'Pclass_3', 'Pclass_1', 'Fare', 'Embarked_C', 'Embarked_S', 'Survived']\n",
      "['Title_Mr', 'Sex_male', 'Sex_female', 'Title_Mrs', 'Title_Miss', 'Pclass_3', 'Pclass_1', 'Fare', 'Embarked_C', 'Embarked_S', 'Survived']\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=44)\n",
    "\n",
    "# Don't actually need these now.\n",
    "# Add y_train to X_train\n",
    "train = pd.concat([X_train, y_train], axis=1)\n",
    "print(list(train))\n",
    "\n",
    "# Add y_test to X_test\n",
    "test = pd.concat([X_test, y_test], axis=1)\n",
    "print(list(test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "596\n",
      "295\n"
     ]
    }
   ],
   "source": [
    "print(len(train))\n",
    "print(len(test)) # saves about 40% for training. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Title_Mr', 'Sex_male', 'Sex_female', 'Title_Mrs', 'Title_Miss', 'Pclass_3', 'Pclass_1', 'Fare', 'Embarked_C', 'Embarked_S']\n",
      "['Title_Mr', 'Sex_male', 'Sex_female', 'Title_Mrs', 'Title_Miss', 'Pclass_3', 'Pclass_1', 'Fare', 'Embarked_C', 'Embarked_S']\n",
      "Int64Index([827, 106, 189, 595, 588, 633, 395, 44, 77, 835], dtype='int64', name='PassengerId')\n"
     ]
    }
   ],
   "source": [
    "print(list(X_train))\n",
    "print(list(X_test))\n",
    "print(X_train.index[0:10])\n",
    "\n",
    "# Passenger ID is the index "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Store the training dataset for later use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#Output this to csv to be read in for predicting values.\n",
    "\n",
    "train_final.to_csv('/home/sophie/projects/Titanic/data/clean_train_53.csv', sep = \" \")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Decision Tree\n",
    "\n",
    "Analysing the different parameters of Decision Trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlMAAADcCAYAAABUOCTMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCiGEEEJoQXSmQgghhBBa8C+ryIFndZuZBQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8fd9d598d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,7))\n",
    "\n",
    "#Max Features\n",
    "plt.subplot(2,3,1) # first of the subplots\n",
    "feature_param = ['auto','sqrt','log2',None]  # testing the keywords that can be given to max_features\n",
    "scores=[]\n",
    "\n",
    "for feature in feature_param:  # loop over each keyword and add result to scores=[]\n",
    "    clf = DecisionTreeClassifier(max_features=feature)  # initiate the model, with a particular number of features\n",
    "    clf.fit(X_train,y_train)  # fit the data to the model\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10) # get the cross validation score for this \n",
    "                                                                          # particular number of features. This is done\n",
    "                                                                          # 10 times (cv = 10)\n",
    "    scores.append(np.mean(scoreCV))  # append the mean of the 10 scores to scores list.\n",
    "\n",
    "plt.plot(scores, '.-')\n",
    "plt.axis('tight')\n",
    "# plt.xlabel('parameter')\n",
    "# plt.ylabel('score')\n",
    "plt.title('Max Features')\n",
    "plt.xticks(range(len(feature_param)), feature_param)\n",
    "plt.grid();\n",
    "\n",
    "#Max Depth\n",
    "plt.subplot(2,3,2)\n",
    "feature_param = range(1,51)\n",
    "scores=[]\n",
    "for feature in feature_param:\n",
    "    clf = DecisionTreeClassifier(max_depth=feature)\n",
    "    clf.fit(X_train,y_train)\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "    scores.append(np.mean(scoreCV))\n",
    "plt.plot(feature_param, scores, '.-')\n",
    "plt.axis('tight')\n",
    "# plt.xlabel('parameter')\n",
    "# plt.ylabel('score')\n",
    "plt.title('Max Depth')\n",
    "plt.grid();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No need to set the maximum number of features. Our initial engineering which narrowed down the section to 10 appears to be better than any of the other keyword options.\n",
    "Setting a maximum depth for the tree of 3 appears to be optimal. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAALEAAACQCAYAAABH2YpWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8fd9cefdd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Min Weight Fraction Leaf\n",
    "plt.subplot(2,3,5)\n",
    "feature_param = np.linspace(0,0.5,10) # create 10 values between 0 and 0.5\n",
    "scores=[]\n",
    "for feature in feature_param:\n",
    "    clf = DecisionTreeClassifier(min_weight_fraction_leaf =feature)\n",
    "    clf.fit(X_train,y_train)\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "    scores.append(np.mean(scoreCV))\n",
    "plt.plot(feature_param, scores, '.-')\n",
    "plt.axis('tight')\n",
    "# plt.xlabel('parameter')\n",
    "# plt.ylabel('score')\n",
    "plt.title('Min Weight Fraction Leaf')\n",
    "plt.grid();\n",
    "\n",
    "#This plot changes everytime. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.          0.05555556  0.11111111  0.16666667  0.22222222  0.27777778\n",
      "  0.33333333  0.38888889  0.44444444  0.5       ]\n"
     ]
    }
   ],
   "source": [
    "print(feature_param)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This changes a lot every time you run the code. Will leave this one out. \n",
    "\n",
    "Now, get the other optimal values, without making the plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "optimal min_samples_split at 14\n",
      "Mean score is 0.820522 with std 0.045472\n",
      "['0.035 ', '0.035 ', '0.034 ', '0.030 ', '0.034 ', '0.032 ', '0.036 ', '0.036 ', '0.037 ', '0.037 ', '0.041 ', '0.044 ', '0.042 ', '0.042 ', '0.045 ', '0.048 ', '0.048 ', '0.049 ', '0.046 ', '0.046 ', '0.048 ', '0.044 ', '0.042 ', '0.044 ', '0.042 ', '0.042 ', '0.040 ', '0.044 ', '0.046 ', '0.048 ', '0.046 ', '0.046 ', '0.046 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.048 ', '0.049 ', '0.050 ', '0.050 ', '0.051 ', '0.051 ', '0.051 ', '0.051 ']\n"
     ]
    }
   ],
   "source": [
    "#Min Samples Split\n",
    "feature_param = range(1,51)\n",
    "scores=[]\n",
    "scores_std=[]\n",
    "\n",
    "scores2=[]\n",
    "for feature in feature_param:\n",
    "    clf = DecisionTreeClassifier(min_samples_split =feature)\n",
    "    clf.fit(X,y) # used the whole training set.\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "    scores.append(np.mean(scoreCV))\n",
    "    scores_std.append(np.std(scoreCV))\n",
    "    \n",
    "    # less recommended method as information from the test set can \"leak\" into the training set, if you tweak the \n",
    "    # parameters.\n",
    "    scores2.append(clf.score(X_test, y_test))\n",
    "    \n",
    "max_index = scores.index(max(scores)) \n",
    "max_index2 = scores2.index(max(scores2))\n",
    "\n",
    "print(\"optimal min_samples_split at %d\" %max_index)\n",
    "\n",
    "print(\"Mean score is %f with std %f\" %(max(scores), scores_std[max_index]))\n",
    "print([\"%.3f \" %s for s in scores_std])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "21\n",
      "8\n"
     ]
    }
   ],
   "source": [
    "# How different are the scores if I use clf.score, or apply cross_validation.cross_val_score?\n",
    "print(max_index)\n",
    "print(max_index2)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Better to use "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "optimal min_samples_leaf at 4\n"
     ]
    }
   ],
   "source": [
    "#Min Samples Leaf\n",
    "feature_param = range(1,51)\n",
    "scores=[]\n",
    "for feature in feature_param:\n",
    "    clf = DecisionTreeClassifier(min_samples_leaf =feature)\n",
    "    clf.fit(X_train,y_train)\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "    scores.append(np.mean(scoreCV))\n",
    "\n",
    "max_index = scores.index(max(scores))    \n",
    "print(\"optimal min_samples_leaf at %d\" %max_index)\n",
    "\n",
    "# If you run this several times it changes between 4 and 10. Maybe try one or the other then switch.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "optimal min_weight_fraction_leaf at 5\n"
     ]
    }
   ],
   "source": [
    "#Min Weight Fraction Leaf\n",
    "feature_param = np.linspace(0,0.5,10)\n",
    "scores=[]\n",
    "for feature in feature_param:\n",
    "    clf = DecisionTreeClassifier(min_weight_fraction_leaf =feature)\n",
    "    clf.fit(X_train,y_train)\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "    scores.append(np.mean(scoreCV))\n",
    "    \n",
    "max_index = scores.index(max(scores))    \n",
    "print(\"optimal min_weight_fraction_leaf at %d\" %max_index)\n",
    "\n",
    "# This changes everytime you do it. Not a stable reliable parameter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "optimal max_leaf_nodes at 6\n"
     ]
    }
   ],
   "source": [
    "#Max Leaf Nodes\n",
    "feature_param = range(2,21)\n",
    "scores=[]\n",
    "for feature in feature_param:\n",
    "    clf = DecisionTreeClassifier(max_leaf_nodes=feature)\n",
    "    clf.fit(X_train,y_train)\n",
    "    scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "    scores.append(np.mean(scoreCV))\n",
    "\n",
    "max_index = scores.index(max(scores))    \n",
    "print(\"optimal max_leaf_nodes at %d\" %max_index)    \n",
    "\n",
    "# Consistently gives 6"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Final precition of score using a combination of the features above"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8159505164\n"
     ]
    }
   ],
   "source": [
    "clf = DecisionTreeClassifier(max_depth = 3, min_samples_split = 14, min_samples_leaf = 4, max_leaf_nodes = 6)\n",
    "clf.fit(X_train,y_train)\n",
    "scoreCV = sklearn.cross_validation.cross_val_score(clf, X, y, cv= 10)\n",
    "score = np.mean(scoreCV)\n",
    "\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### So, using Decision Trees with very specific parameters I can get a score up to 0.816? Nope, not that high."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}