from sklearn.model_selection import StratifiedKFold
from sklearn import preprocessing
from sklearn.linear_model import LogisticRegression, Ridge
from sklearn.preprocessing import LabelBinarizer
from stacked_generalization.lib.stacking import StackedClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.neighbors import KNeighborsClassifier
import pandas as pd
import numpy as np
import re
class DataReader(object):
def __init__(self, file_name):
self.file_name = file_name
def disp_hist(self, data, label, bins):
temp = [i[label].dropna() for i in data]
plt.hist(temp, histtype='barstacked', bins=bins)
plt.show()
def pre_process(self, drop=True, title_to_onehot=True, norm_fare=True):
def get_title(name):
title_search = re.search(' ([A-Za-z]+)\.', name)
if title_search:
return title_search.group(1)
return ""
def normalize_fare(data):
new_data = None
for embarked in (0, 1, 2):
temp = data[data.Embarked == embarked]
temp['Fare'] /= temp['Fare'].values.mean()
if new_data is None:
new_data = temp
else:
new_data = pd.concat([new_data, temp])
new_data = new_data.sort('PassengerId')
return new_data
data = pd.read_csv(self.file_name).replace('male',0).replace('female',1)
data['Age'].fillna(data.Age.median(), inplace=True)
data['Fare'].fillna(data.Fare.median(), inplace=True)
data['FamilySize'] = data['SibSp'] + data['Parch'] + 1
data['Embarked'] = data['Embarked'].replace('S',0).replace('C',1).replace('Q',2)
data['Embarked'].fillna(0, inplace=True)
if norm_fare:
data = normalize_fare(data)
# Get all the titles and print how often each one occurs.
titles = data["Name"].apply(get_title)
print(pd.value_counts(titles))
# Map each title to an integer. Some titles are very rare, and are compressed into the same codes as other titles.
title_mapping = {"Dona": 1, "Mr": 1, "Miss": 2, "Mrs": 3, "Master": 4, "Dr": 5, "Rev": 6, "Major": 7, "Col": 7, "Mlle": 8, "Mme": 8, "Don": 9, "Lady": 10, "Countess": 10, "Jonkheer": 10, "Sir": 9, "Capt": 7, "Ms": 2}
for k,v in title_mapping.items():
titles[titles == k] = v
# Add in the title column.
data['Title'] = titles
data['Title'].fillna(1, inplace=True)
#data['Pos'] = data["Title"] + data['Pclass']
if drop:
#data = data.drop(['Name', 'SibSp', 'Parch', 'Ticket', 'Pclass', 'Cabin', 'Embarked'], axis=1)
data = data.drop(['Name', 'SibSp', 'Parch', 'Ticket', 'Cabin', 'Embarked'], axis=1)
#data = data.drop(['Name', 'SibSp', 'Parch', 'Ticket', 'Cabin', 'Embarked', 'Pclass', 'Title'], axis=1)
print(data.keys())
if title_to_onehot:
self.encode(data, 'Title', [i for i in range(1, 11)])
data = data.drop(['Title'], axis=1)
return data
def encode(self, data, label, value_set=None):
le =LabelBinarizer()
if value_set is None:
encoded = le.fit_transform(data[label])
else:
le.fit(value_set)
encoded = le.transform(data[label])
for i in range(encoded.shape[1]):
new_label = '{0}_is_{1}'.format(label, i)
data[new_label] = encoded[:,i]
def split_by_label(self, data, label='Survived'):
split_data = []
for element in set(data[label]):
split_data.append(data[data[label]==element])
return split_data
def get_sample(self, N=600, scale=False):
all_data = self.pre_process(self.file_name)
#print('data_type: ' + str(all_data.dtypes))
all_data = all_data.values
xs = all_data[:, 2:]
y = all_data[:, 1]
if scale:
xs = preprocessing.scale(xs)
if N != -1:
perm = np.random.permutation(xs.shape[0])
xs = xs[perm]
y = y[perm]
xs_train, xs_test = np.split(xs, [N])
y_train, y_test = np.split(y, [N])
return xs_train, xs_test, y_train, y_test
else:
return xs, y
def summarize_about_same_ticket(self, data):
data = data.drop(['PassengerId', 'Name', 'SibSp', 'Parch', 'Cabin', 'FamilySize'], axis=1)
for num in data[data.Age <= 5.0]['Ticket']:
print('num:' + num)
print(data[data.Ticket == num])
class TestDataReader(DataReader):
def get_sample(self, N=-1):
all_data = self.pre_process(self.file_name)
all_data = all_data.values
xs = all_data[:, 1:]
pid = all_data[:, 0]
return pid, xs
def write_result(pid, output, suffix=''):
import csv
import datetime
suffix += datetime.datetime.today().strftime("%Y-%m-%d-%H-%M-%S")
with open("predict_result_data_{0}.csv".format(suffix), "w") as f:
writer = csv.writer(f, lineterminator='\n')
writer.writerow(["PassengerId", "Survived"])
for pid, survived in zip(pid.astype(int), output.astype(int)):
writer.writerow([pid, survived])
if __name__ == '__main__':
import os
if not os.path.isfile('train.csv'):
raise Exception('This example is data analysis for Kaggle Titanic Competition.' +
'For trying this example, you should download "train.csv" from https://www.kaggle.com/c/titanic.')
train = True
full_cv = True
test = False
train_dr = DataReader('train.csv')
bclf = LogisticRegression(random_state=1)
clfs = [
RandomForestClassifier(n_estimators=50, criterion = 'gini', random_state=1),
ExtraTreesClassifier(n_estimators=50, criterion = 'gini', random_state=1),
ExtraTreesClassifier(n_estimators=50, criterion = 'gini', random_state=2),
GradientBoostingClassifier(n_estimators=25, random_state=1),
GradientBoostingClassifier(n_estimators=40, random_state=1),
Ridge(random_state=1),
#KNeighborsClassifier(n_neighbors=4)
#LogisticRegression(random_state=1)
]
sl = StackedClassifier(bclf, clfs, n_folds=3, verbose=2)
#fsl = FWSLClassifier(bclf, clfs, feature=xs_train[:, 0])
if train:# evalute by hold-out and out-of-bugs
sl = StackedClassifier(bclf, clfs, n_folds=3, verbose=2, oob_score_flag=True)
xs_train, xs_test, y_train, y_test = train_dr.get_sample()
sl.fit(xs_train, y_train)
score = sl.score(xs_test, y_test)
print('score: {0}'.format(score))
print('oob_score: {0}'.format(sl.oob_score_))
if full_cv: #cross validation
sl = StackedClassifier(bclf, clfs, oob_score_flag=False,verbose=2)
xs_train, y_train = train_dr.get_sample(-1)
score = []
for train_index, test_index in StratifiedKFold(3).split(xs_train, y_train):
sl.fit(xs_train[train_index], y_train[train_index])
score.append(sl.score(xs_train[test_index], y_train[test_index]))
print('full-cv score: {0}'.format(score))
if test: #to make pb leader board data.
xs_train, y_train = train_dr.get_sample(-1)
sl.fit(xs_train, y_train)
test_dr = TestDataReader('test.csv')
pid, xs_test = test_dr.get_sample(-1)
output = sl.predict(xs_test)
write_result(pid, output, sl.tostr())
