{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Kaggle's Titanic Starter Competition: first attempt\n", "\n", "Quick and dirty solution(s) for Kaggle's starter competition [Titanic: Machine Learning from Disaster](https://www.kaggle.com/c/titanic).\n", "\n", "In this notebook my goal was to get to a submission as quickly as possible, and as such I've skipped exploratory analysis and dropped a couple of variables that might be useful. But by making some quick decisions on how to preprocess data and applying off the shelf models, we could compare results and attempt a couple of submissions, getting up to 0.736 accuracy. From what I read it should be possible to get closer to 0.8, but this is a decent start.\n", "\n", "## Data pre-processing\n", "\n", "We need to load in the training data set and decide on pre-processing steps including filling in missing data, normalizing values etc.\n", "\n", "```\n", "VARIABLE DESCRIPTIONS:\n", "survival Survival\n", " (0 = No; 1 = Yes)\n", "pclass Passenger Class\n", " (1 = 1st; 2 = 2nd; 3 = 3rd)\n", "name Name\n", "sex Sex\n", "age Age\n", "sibsp Number of Siblings/Spouses Aboard\n", "parch Number of Parents/Children Aboard\n", "ticket Ticket Number\n", "fare Passenger Fare\n", "cabin Cabin\n", "embarked Port of Embarkation\n", " (C = Cherbourg; Q = Queenstown; S = Southampton)\n", "```\n", "\n", "### Loading the data" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdSurvivedPclassNameSexAgeSibSpParchTicketFareCabinEmbarked
0103Braund, Mr. Owen Harrismale2210A/5 211717.2500NaNS
1211Cumings, Mrs. John Bradley (Florence Briggs Th...female3810PC 1759971.2833C85C
2313Heikkinen, Miss. Lainafemale2600STON/O2. 31012827.9250NaNS
3411Futrelle, Mrs. Jacques Heath (Lily May Peel)female351011380353.1000C123S
4503Allen, Mr. William Henrymale35003734508.0500NaNS
5603Moran, Mr. JamesmaleNaN003308778.4583NaNQ
6701McCarthy, Mr. Timothy Jmale54001746351.8625E46S
7803Palsson, Master. Gosta Leonardmale23134990921.0750NaNS
8913Johnson, Mrs. Oscar W (Elisabeth Vilhelmina Berg)female270234774211.1333NaNS
91012Nasser, Mrs. Nicholas (Adele Achem)female141023773630.0708NaNC
\n", "
" ], "text/plain": [ " PassengerId Survived Pclass \\\n", "0 1 0 3 \n", "1 2 1 1 \n", "2 3 1 3 \n", "3 4 1 1 \n", "4 5 0 3 \n", "5 6 0 3 \n", "6 7 0 1 \n", "7 8 0 3 \n", "8 9 1 3 \n", "9 10 1 2 \n", "\n", " Name Sex Age SibSp \\\n", "0 Braund, Mr. Owen Harris male 22 1 \n", "1 Cumings, Mrs. John Bradley (Florence Briggs Th... female 38 1 \n", "2 Heikkinen, Miss. Laina female 26 0 \n", "3 Futrelle, Mrs. Jacques Heath (Lily May Peel) female 35 1 \n", "4 Allen, Mr. William Henry male 35 0 \n", "5 Moran, Mr. James male NaN 0 \n", "6 McCarthy, Mr. Timothy J male 54 0 \n", "7 Palsson, Master. Gosta Leonard male 2 3 \n", "8 Johnson, Mrs. Oscar W (Elisabeth Vilhelmina Berg) female 27 0 \n", "9 Nasser, Mrs. Nicholas (Adele Achem) female 14 1 \n", "\n", " Parch Ticket Fare Cabin Embarked \n", "0 0 A/5 21171 7.2500 NaN S \n", "1 0 PC 17599 71.2833 C85 C \n", "2 0 STON/O2. 3101282 7.9250 NaN S \n", "3 0 113803 53.1000 C123 S \n", "4 0 373450 8.0500 NaN S \n", "5 0 330877 8.4583 NaN Q \n", "6 0 17463 51.8625 E46 S \n", "7 1 349909 21.0750 NaN S \n", "8 2 347742 11.1333 NaN S \n", "9 0 237736 30.0708 NaN C " ] }, "execution_count": 1, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import pandas as pd\n", "\n", "training_data = pd.read_csv('train.csv')\n", "\n", "training_data.head(10)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Finding features with missing data\n", "\n", "Looks like there are plenty of columns with missing data." ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Total rows: 10692\n", "Variables with missing rows:\n" ] }, { "data": { "text/plain": [ "PassengerId 0\n", "Survived 0\n", "Pclass 0\n", "Name 0\n", "Sex 0\n", "Age 177\n", "SibSp 0\n", "Parch 0\n", "Ticket 0\n", "Fare 0\n", "Cabin 687\n", "Embarked 2\n", "dtype: int64" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "print(\"Total rows: {}\".format(training_data.size))\n", "\n", "print(\"Variables with missing rows:\")\n", "training_data.isnull().sum()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Handling missing data, mapping categorical data and scaling \n", "\n", "Things we need to take care of in preprocessing data:\n", "\n", "- Selecting features: after eyeballing the data, we drop a couple of columns that seem like a pain to deal with and/or are missing many values.\n", "- Missing values: it appears the only missing values (among features we're working with) are for quantitative variables. For those values, we'll replace with median value for that column.\n", "- Categorical data: we map binary variables to 0, 1\n", "- Scaling:\n", " - categorical values: we push to -1, 1\n", " - quantitative: we use scikit-learn's standard scalar to center the feature columns at mean 0 with standard deviation 1 so that the feature columns take the form of a normal distribution\n", " - note: we provide an option to not scale as it isn't necessary for tree based models\n", " \n", "The helper method I developed to perform this scaling is below\n" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# %load titanic.py\n", "from sklearn.preprocessing import StandardScaler\n", "\n", "\n", "def make_preprocesser(training_data):\n", " \"\"\"\n", " Constructs a preprocessing function ready to apply to new dataframes.\n", "\n", " Crucially, the interpolating that is done based on the training data set\n", " is remembered so it can be applied to test datasets (e.g the mean age that\n", " is used to fill in missing values for 'Age' will be fixed based on the mean\n", " age within the training data set).\n", "\n", " Summary by column:\n", "\n", " ['PassengerId',\n", " 'Survived', # this is our target, not a feature\n", " 'Pclass', # keep as is: ordinal value should work, even though it's inverted (higher number is lower class cabin)\n", " 'Name', # omit (could try some fancy stuff like inferring ethnicity, but skip for now)\n", " 'Sex', # code to 0 / 1\n", " 'Age', # replace missing with median\n", " 'SibSp',\n", " 'Parch',\n", " 'Ticket', # omit (doesn't seem like low hanging fruit, could look more closely for pattern later)\n", " 'Fare', # keep, as fare could be finer grained proxy for socio economic status, sense of entitlement / power in getting on boat\n", " 'Cabin', # omit (10% are missing values, could look more closely later, one idea would be to break this out into a few different boolean variables for major section, A->E)\n", " 'Embarked'] # omit (could one-hot encode it, but can't see how this would affect survivorship, let's be lazy to start)\n", "\n", " Params:\n", " df: pandas.DataFrame containing the training data\n", " Returns:\n", " fn: a function to preprocess a dataframe (either before training or fitting a new dataset)\n", " \"\"\"\n", "\n", " def pick_features(df):\n", " return df[['PassengerId', 'Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare']]\n", "\n", " # save median Age so we can use it to fill in missing data consistently\n", " # on any dataset\n", " median_age_series = training_data[['Age', 'Fare']].median()\n", "\n", " def fix_missing(df):\n", " return df.fillna(median_age_series)\n", "\n", " def map_categorical(df):\n", " df['Sex'] = df['Sex'].map({'male': 0, 'female': 1})\n", " return df\n", "\n", " # We want standard scaling fit on the training data, so we get a scaler ready\n", " # for application now. It needs to be applied to data that already has the other\n", " # pre-processing applied.\n", " training_data_all_but_scaled = map_categorical(fix_missing(pick_features(training_data)))\n", " stdsc = StandardScaler()\n", " stdsc.fit(training_data_all_but_scaled[['Pclass', 'Age', 'SibSp', 'Parch', 'Fare']])\n", "\n", " def scale_df(df):\n", " df[['Pclass', 'Age', 'SibSp', 'Parch', 'Fare']] = \\\n", " stdsc.transform(df[['Pclass', 'Age', 'SibSp', 'Parch', 'Fare']])\n", " df[['Sex']] = df[['Sex']].applymap(lambda x: 1 if x == 1 else -1)\n", " return df\n", "\n", " def preprocess(df, scale=True):\n", " \"\"\"\n", " Preprocesses a dataframe so it is ready for use with a model (either for training or prediction).\n", "\n", " Params:\n", " scale: whether to apply feature scaling. E.g with random forests feature scaling isn't necessary.\n", " \"\"\"\n", " all_but_scaled = map_categorical(fix_missing(pick_features(df)))\n", " if scale:\n", " return scale_df(all_but_scaled)\n", " else:\n", " return all_but_scaled\n", "\n", " return preprocess\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Applying off the shelf classifiers on the way to quick first submission\n", "\n", "Let's try out a few classifiers on test vs training portions of the dataset, using our preprocessor function as appropriate.\n", "\n", "Note that we split the training set provided into a training and test set so we can get some insight into how well each model might generalize to unseen data." ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdPclassSexAgeSibSpParchFare
013022107.2500
1211381071.2833
233126007.9250
3411351053.1000
453035008.0500
563028008.4583
6710540051.8625
783023121.0750
8931270211.1333
91021141030.0708
\n", "
" ], "text/plain": [ " PassengerId Pclass Sex Age SibSp Parch Fare\n", "0 1 3 0 22 1 0 7.2500\n", "1 2 1 1 38 1 0 71.2833\n", "2 3 3 1 26 0 0 7.9250\n", "3 4 1 1 35 1 0 53.1000\n", "4 5 3 0 35 0 0 8.0500\n", "5 6 3 0 28 0 0 8.4583\n", "6 7 1 0 54 0 0 51.8625\n", "7 8 3 0 2 3 1 21.0750\n", "8 9 3 1 27 0 2 11.1333\n", "9 10 2 1 14 1 0 30.0708" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "preprocess = make_preprocesser(training_data)\n", "\n", "td_preprocessed_unscaled = preprocess(training_data, scale=False)\n", "\n", "td_preprocessed_unscaled.head(10)" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdPclassSexAgeSibSpParchFare
010.827377-1-0.5657360.432793-0.473674-0.502445
12-1.56610710.6638610.432793-0.4736740.786845
230.8273771-0.258337-0.474545-0.473674-0.488854
34-1.56610710.4333120.432793-0.4736740.420730
450.827377-10.433312-0.474545-0.473674-0.486337
560.827377-1-0.104637-0.474545-0.473674-0.478116
67-1.566107-11.893459-0.474545-0.4736740.395814
780.827377-1-2.1027332.2474700.767630-0.224083
890.8273771-0.181487-0.4745452.008933-0.424256
910-0.3693651-1.1805350.432793-0.473674-0.042956
\n", "
" ], "text/plain": [ " PassengerId Pclass Sex Age SibSp Parch Fare\n", "0 1 0.827377 -1 -0.565736 0.432793 -0.473674 -0.502445\n", "1 2 -1.566107 1 0.663861 0.432793 -0.473674 0.786845\n", "2 3 0.827377 1 -0.258337 -0.474545 -0.473674 -0.488854\n", "3 4 -1.566107 1 0.433312 0.432793 -0.473674 0.420730\n", "4 5 0.827377 -1 0.433312 -0.474545 -0.473674 -0.486337\n", "5 6 0.827377 -1 -0.104637 -0.474545 -0.473674 -0.478116\n", "6 7 -1.566107 -1 1.893459 -0.474545 -0.473674 0.395814\n", "7 8 0.827377 -1 -2.102733 2.247470 0.767630 -0.224083\n", "8 9 0.827377 1 -0.181487 -0.474545 2.008933 -0.424256\n", "9 10 -0.369365 1 -1.180535 0.432793 -0.473674 -0.042956" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "td_preprocessed = preprocess(training_data, scale=True)\n", "\n", "td_preprocessed.head(10)" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Decision Tree training/test accuracy: 0.84 | 0.82\n", "Random forest training/test accuracy: 0.96 | 0.82\n", "Logistic Regression training/test accuracy: 0.79 | 0.80\n", "Linear SVM training/test accuracy: 0.79 | 0.79\n", "Kernel SVM training/test accuracy: 0.84 | 0.81\n" ] } ], "source": [ "from sklearn.tree import DecisionTreeClassifier\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.svm import SVC\n", "\n", "\n", "from sklearn.metrics import accuracy_score\n", "from sklearn.cross_validation import train_test_split\n", "\n", "for name, should_scale, model in [\n", " (\"Decision Tree\", False, DecisionTreeClassifier(criterion='entropy', max_depth=3, random_state=0)),\n", " (\"Random forest\", False, RandomForestClassifier(criterion='entropy',\n", " n_estimators=10, \n", " random_state=1,\n", " n_jobs=2)),\n", " (\"Logistic Regression\", True, LogisticRegression(C=100.0, random_state=0)),\n", " (\"Linear SVM\", True, SVC(kernel='linear', C=1.0, random_state=0)),\n", " (\"Kernel SVM\", True, SVC(kernel='rbf', random_state=0, gamma=0.10, C=10.0))]:\n", " pre_processed = td_preprocessed if should_scale else td_preprocessed_unscaled\n", " X = pre_processed.loc[:,'Pclass':]\n", " y = training_data['Survived']\n", "\n", " X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=0)\n", "\n", " model.fit(X_train, y_train)\n", " print(\"{} training/test accuracy: {:.2f} | {:.2f}\".format(\n", " name, \n", " accuracy_score(y_train, model.predict(X_train)),\n", " accuracy_score(y_test, model.predict(X_test))))\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The biggest delta between training and test is with random forests which suggests that when trained it's apparent advantage could be lost due to overfitting.\n", "\n", "## First submissions to Kaggle\n", "\n", "Let's make our first submissions. We can re-train the most promising off the shelf classifier, random forest, on the entire training data set, run it on kaggle's test data set and submit." ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Random forest full training set accuracy: 0.96\n" ] }, { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdPclassSexAgeSibSpParchFare
08923034.5007.8292
18933147.0107.0000
28942062.0009.6875
38953027.0008.6625
48963122.01112.2875
\n", "
" ], "text/plain": [ " PassengerId Pclass Sex Age SibSp Parch Fare\n", "0 892 3 0 34.5 0 0 7.8292\n", "1 893 3 1 47.0 1 0 7.0000\n", "2 894 2 0 62.0 0 0 9.6875\n", "3 895 3 0 27.0 0 0 8.6625\n", "4 896 3 1 22.0 1 1 12.2875" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "fully_trained_forest = RandomForestClassifier(criterion='entropy',\n", " n_estimators=10, \n", " random_state=1,\n", " n_jobs=2)\n", "X = td_preprocessed_unscaled.loc[:,'Pclass':]\n", "y = training_data['Survived']\n", "fully_trained_forest.fit(X, y)\n", "print(\"Random forest full training set accuracy: {:.2f}\".format(accuracy_score(y, fully_trained_forest.predict(X))))\n", "\n", "test_data = pd.read_csv('test.csv')\n", "test_data_preprocessed = preprocess(test_data, scale=False)\n", "test_data_preprocessed.head()" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdSurvived
08920
18930
28941
38951
48960
\n", "
" ], "text/plain": [ " PassengerId Survived\n", "0 892 0\n", "1 893 0\n", "2 894 1\n", "3 895 1\n", "4 896 0" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "test_data_predict = fully_trained_forest.predict(test_data_preprocessed.loc[:,'Pclass':])\n", "\n", "submission_df = test_data_preprocessed[['PassengerId']].copy()\n", "submission_df['Survived'] = test_data_predict\n", "\n", "submission_df.to_csv(\n", " 'titanic_submission_random_forest_1.csv',\n", " index=False)\n", "\n", "submission_df.head()" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABPoAAADyCAYAAADUZqvSAAAMGWlDQ1BJQ0MgUHJvZmlsZQAASImV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\nAAAAAElFTkSuQmCC\n", "text/plain": [ "" ] }, "execution_count": 9, "metadata": { "image/png": { "width": 600 } }, "output_type": "execute_result" } ], "source": [ "from IPython.display import Image\n", "Image(filename='kaggle-submission-random-forest.png', width=600) " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "It's interesting that we only acheived 73% accuracy when our local test set was 82%.\n", "\n", "Before digging deeper, let's also try a submission using logistic regression." ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Logistic Regression full training set accuracy: 0.79\n" ] }, { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdPclassSexAgeSibSpParchFare
08920.827377-10.394887-0.474545-0.473674-0.490783
18930.82737711.3555100.432793-0.473674-0.507479
2894-0.369365-12.508257-0.474545-0.473674-0.453367
38950.827377-1-0.181487-0.474545-0.473674-0.474005
48960.8273771-0.5657360.4327930.767630-0.401017
\n", "
" ], "text/plain": [ " PassengerId Pclass Sex Age SibSp Parch Fare\n", "0 892 0.827377 -1 0.394887 -0.474545 -0.473674 -0.490783\n", "1 893 0.827377 1 1.355510 0.432793 -0.473674 -0.507479\n", "2 894 -0.369365 -1 2.508257 -0.474545 -0.473674 -0.453367\n", "3 895 0.827377 -1 -0.181487 -0.474545 -0.473674 -0.474005\n", "4 896 0.827377 1 -0.565736 0.432793 0.767630 -0.401017" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "fully_trained_lgr = LogisticRegression(C=100.0, random_state=0)\n", "\n", "X_scaled = td_preprocessed.loc[:,'Pclass':]\n", "y = training_data['Survived']\n", "fully_trained_lgr.fit(X_scaled, y)\n", "print(\"Logistic Regression full training set accuracy: {:.2f}\".format(\n", " accuracy_score(y, fully_trained_lgr.predict(X_scaled))))\n", "\n", "test_data = pd.read_csv('test.csv')\n", "test_data_preprocessed_scaled = preprocess(test_data, scale=True)\n", "test_data_preprocessed_scaled.head()" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PassengerIdSurvived
08920
18930
28940
38950
48961
\n", "
" ], "text/plain": [ " PassengerId Survived\n", "0 892 0\n", "1 893 0\n", "2 894 0\n", "3 895 0\n", "4 896 1" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "test_data_predict = fully_trained_lgr.predict(test_data_preprocessed_scaled.loc[:,'Pclass':])\n", "\n", "submission_df = test_data_preprocessed_scaled[['PassengerId']].copy()\n", "submission_df['Survived'] = test_data_predict\n", "\n", "submission_df.to_csv(\n", " 'titanic_submission_logistic_regression_1.csv',\n", " index=False)\n", "\n", "submission_df.head()" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABOgAAADsCAYAAADO4ZnsAAAMGWlDQ1BJQ0MgUHJvZmlsZQAASImV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\nkAAJkAAJkAAJkAAJkEC1Cfx/croLSTHJ1jMAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "execution_count": 12, "metadata": { "image/png": { "width": 600 } }, "output_type": "execute_result" } ], "source": [ "Image(filename='kaggle-submission-logistic-regression.png', width=600) " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As earlier suspected, any advantage the random forest model had over logistic regression was due to overfitting and didn't generalize to perform any better once submitted to Kaggle. Logistic regression performed about the same." ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.4.4" } }, "nbformat": 4, "nbformat_minor": 0 }