{ "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "474cb36d-9bd4-6878-e2f5-c83178862f29" }, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "10c21343-5962-75cb-7d19-d4fff5bb884f" }, "outputs": [], "source": [ "# Loading datasets\n", "train_df = pd.read_csv('../input/train.csv')\n", "test_df = pd.read_csv('../input/test.csv')\n", "combine = [train_df, test_df]\n", "\n", "passenger_ids = test_df['PassengerId']" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "e6546039-1fee-cd93-3896-1adc33308ad9" }, "outputs": [], "source": [ "# Estimating missing values\n", "# Feature: Age\n", "guess_age_df = train_df[['Sex', 'Pclass', 'Age']].groupby(['Sex', 'Pclass']).median()\n", "guess_age_df = guess_age_df.reset_index()\n", "\n", "for dataset in combine:\n", " for sex in ['male', 'female']:\n", " for pclass in [1, 2, 3]:\n", " dataset.loc[ (dataset['Age'].isnull()) & (dataset['Sex'] == sex) & \\\n", " (dataset['Pclass'] == pclass), 'Age' ] = \\\n", " guess_age_df.loc[ (guess_age_df['Sex'] == sex) & \\\n", " (guess_age_df['Pclass'] == pclass), 'Age'].values[0]\n", "\n", "# Feature: Embarked\n", "guess_embarked = train_df['Embarked'].mode()[0]\n", "for dataset in combine:\n", " dataset['Embarked'].fillna( guess_embarked, inplace=True )\n", "\n", "# Feature: Fare\n", "guess_fare = train_df['Fare'].median()\n", "for dataset in combine:\n", " dataset['Fare'].fillna( guess_fare, inplace=True )" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "2b7f7af0-f9f9-de20-37bd-1cea2df4fb4f" }, "outputs": [], "source": [ "# Obtaining new features\n", "# Title\n", "for dataset in combine:\n", " # creates it\n", " dataset['Title'] = dataset['Name'].str.extract(', (\\w+)\\.', expand=True)\n", "\n", " # replace some values\n", " dataset['Title'] = dataset['Title'].replace(['Capt', 'Col', 'Don', 'Dr', 'Jonkheer',\\\n", " 'Lady', 'Major', 'Rev', 'Sir'], 'Rare')\n", " dataset['Title'] = dataset['Title'].replace(['Mlle', 'Ms'], 'Miss')\n", " dataset['Title'] = dataset['Title'].replace('Mme', 'Mrs')\n", "\n", " # transform it to numerical\n", " dataset['Title'] = dataset['Title'].map({'Master': 1, 'Miss': 2, 'Mr': 3, \\\n", " 'Mrs': 4, 'Rare': 5})\n", " dataset['Title'] = dataset['Title'].fillna(0).astype(int)\n", "\n", "# AgeBand\n", "for dataset in combine:\n", " dataset['AgeBand'] = pd.cut(train_df['Age'], 5)\n", " \n", " # discretizing\n", " for index, value in enumerate(train_df['AgeBand'].unique()):\n", " dataset.loc[ dataset['AgeBand'] == value, 'Age' ] = index\n", "\n", "# Creating new feature: FamilySize\n", "for dataset in combine:\n", " dataset['FamilySize'] = dataset['SibSp'] + dataset['Parch'] + 1\n", " \n", "# Discretize continuous features: Age\n", "for dataset in combine:\n", " dataset['FareBand'] = pd.cut(train_df['Fare'], 4)\n", " \n", " # discretizing\n", " for index, value in enumerate(train_df['FareBand'].unique()):\n", " dataset.loc[ dataset['FareBand'] == value, 'Fare' ] = index" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "bb0b4919-2e6f-d2a3-036a-5362450eb613" }, "outputs": [], "source": [ "# Dropping features\n", "drop_columns = ['Ticket', 'Cabin', 'PassengerId', 'Name', 'AgeBand', 'SibSp', 'Parch', 'FareBand']\n", "\n", "train_df = train_df.drop(drop_columns, axis=1)\n", "test_df = test_df.drop(drop_columns, axis=1)\n", "\n", "combine = [train_df, test_df]" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "e6676c31-0548-c6dd-3772-64a20d3ccffe" }, "outputs": [], "source": [ "print(test_df['Fare'].isnull().any())\n", "\n", "train_df.head()" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "74df3948-7274-a02b-125f-80702640178b" }, "outputs": [], "source": [ "# Mapping features to numerical\n", "for dataset in combine:\n", " # Feature: Sex\n", " dataset['Sex'] = dataset['Sex'].map({'male': 0, 'female': 1})\n", " # Embarked\n", " dataset['Embarked'] = dataset['Embarked'].map({'C': 0, 'Q': 1, 'S': 2})\n", " \n", "for dataset in combine:\n", " dataset['Sex'] = dataset['Sex'].astype(int)\n", " dataset['Age'] = dataset['Embarked'].astype(int)\n", " dataset['Age'] = dataset['Age'].astype(int)\n", " dataset['Fare'] = dataset['Fare'].astype(int)\n", " dataset['Title'] = dataset['Title'].astype(int)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "0e673fec-4edf-ea83-3926-60f632ba3f8b" }, "outputs": [], "source": [ "from sklearn.cross_validation import train_test_split\n", "\n", "train, test = train_test_split(train_df, train_size=0.8)\n", "\n", "X_train = train.drop('Survived', axis=1)\n", "y_train = train['Survived']\n", "X_test = test.drop('Survived', axis=1)\n", "y_test = test['Survived']" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "91e2a129-506d-869e-ed02-28d327c125e5" }, "outputs": [], "source": [ "from sklearn.ensemble import RandomForestClassifier\n", "\n", "random_forest = RandomForestClassifier()\n", "random_forest.fit(X_train, y_train)\n", "random_forest.score(X_test, y_test)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "_cell_guid": "5e3b6823-cdd6-2f62-c347-5243e336d9f7" }, "outputs": [], "source": [ "final_clf = RandomForestClassifier()\n", "final_clf.fit(train_df.drop('Survived', axis=1), train_df['Survived'])\n", "predictions = final_clf.predict(test_df)\n", "\n", "submission_df = pd.DataFrame({ 'PassengerId': passenger_ids, 'Survived': predictions})\n", "submission_df" ] } ], "metadata": { "_change_revision": 0, "_is_fork": false, "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": "3.6.0" } }, "nbformat": 4, "nbformat_minor": 0 }