-
Notifications
You must be signed in to change notification settings - Fork 0
/
titanic.py
188 lines (141 loc) · 8.2 KB
/
titanic.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
# -*- coding: utf-8 -*-
"""
Created on Tue Jan 22 05:43:40 2019
@author: KIIT
"""
#titainc problem
import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestClassifier, ExtraTreesRegressor
from sklearn import model_selection
import re
import operator
from sklearn.feature_selection import SelectKBest, f_classif
#Print you can execute arbitrary python code
train = pd.read_csv("train.csv", dtype={"Age": np.float64}, )
test = pd.read_csv("test.csv", dtype={"Age": np.float64}, )
target = train["Survived"].values
full = pd.concat([train, test])
#print(full.head())
#print(full.describe())
#print(full.info())
full['surname'] = full["Name"].apply(lambda x: x.split(',')[0].lower())
full["Title"] = full["Name"].apply(lambda x: re.search(' ([A-Za-z]+)\.',x).group(1))
title_mapping = {"Mr": 1, "Miss": 2, "Mrs": 3, "Master": 4, "Dr": 5, "Rev": 6, "Major": 7, "Col": 7, "Mlle": 2, "Mme": 3,"Don": 9,"Dona": 9, "Lady": 10, "Countess": 10, "Jonkheer": 10, "Sir": 9, "Capt": 7, "Ms": 2}
full["TitleCat"] = full.loc[:,'Title'].map(title_mapping)
full["FamilySize"] = full["SibSp"] + full["Parch"] + 1
full["FamilySize"] = pd.cut(full["FamilySize"], bins=[0,1,4,20], labels=[0,1,2])
full["NameLength"] = full["Name"].apply(lambda x: len(x))
full["Embarked"] = pd.Categorical.from_array(full.Embarked).codes
full["Fare"] = full["Fare"].fillna(8.05)
full = pd.concat([full,pd.get_dummies(full['Sex'])],axis=1)
full['CabinCat'] = pd.Categorical.from_array(full.Cabin.fillna('0').apply(lambda x: x[0])).codes
# function to get oven/odd/null from cabine
def get_type_cabine(cabine):
# Use a regular expression to search for a title.
cabine_search = re.search('\d+', cabine)
# If the title exists, extract and return it.
if cabine_search:
num = cabine_search.group(0)
if np.float64(num) % 2 == 0:
return '2'
else:
return '1'
return '0'
full["Cabin"] = full["Cabin"].fillna(" ")
full["CabinType"] = full["Cabin"].apply(get_type_cabine)
#print(pd.value_counts(full["CabinType"]))
#### CHILD/FEMALE ADULT/MALE ADULT------------------------------------------------------------
child_age = 18
def get_person(passenger):
age, sex = passenger
if (age < child_age):
return 'child'
elif (sex == 'female'):
return 'female_adult'
else:
return 'male_adult'
full = pd.concat([full, pd.DataFrame(full[['Age', 'Sex']].apply(get_person, axis=1), columns=['person'])],axis=1)
full = pd.concat([full,pd.get_dummies(full['person'])],axis=1)
### FEATURES BASED ON TICKET --------------------------------------------------------
table_ticket = pd.DataFrame(full["Ticket"].value_counts())
table_ticket.rename(columns={'Ticket':'Ticket_Members'}, inplace=True)
table_ticket['Ticket_perishing_women'] = full.Ticket[(full.female_adult == 1.0)
& (full.Survived == 0.0)
& ((full.Parch > 0) | (full.SibSp > 0))].value_counts()
table_ticket['Ticket_perishing_women'] = table_ticket['Ticket_perishing_women'].fillna(0)
table_ticket['Ticket_perishing_women'][table_ticket['Ticket_perishing_women'] > 0] = 1.0
table_ticket['Ticket_surviving_men'] = full.Ticket[(full.male_adult == 1.0)
& (full.Survived == 1.0)
& ((full.Parch > 0) | (full.SibSp > 0))].value_counts()
table_ticket['Ticket_surviving_men'] = table_ticket['Ticket_surviving_men'].fillna(0)
table_ticket['Ticket_surviving_men'][table_ticket['Ticket_surviving_men'] > 0] = 1.0
table_ticket["Ticket_Id"]= pd.Categorical.from_array(table_ticket.index).codes
# compress under 3 members into one code.
table_ticket["Ticket_Id"][table_ticket["Ticket_Members"] < 3 ] = -1
table_ticket["Ticket_Members"] = pd.cut(table_ticket["Ticket_Members"], bins=[0,1,4,20], labels=[0,1,2])
full = pd.merge(full, table_ticket, left_on="Ticket",right_index=True,how='left', sort=False)
### FEATURES BASED ON SURNAME --------------------------------------------------------
table_surname = pd.DataFrame(full["surname"].value_counts())
table_surname.rename(columns={'surname':'Surname_Members'}, inplace=True)
table_surname['Surname_perishing_women'] = full.surname[(full.female_adult == 1.0)
& (full.Survived == 0.0)
& ((full.Parch > 0) | (full.SibSp > 0))].value_counts()
table_surname['Surname_perishing_women'] = table_surname['Surname_perishing_women'].fillna(0)
table_surname['Surname_perishing_women'][table_surname['Surname_perishing_women'] > 0] = 1.0
table_surname['Surname_surviving_men'] = full.surname[(full.male_adult == 1.0)
& (full.Survived == 1.0)
& ((full.Parch > 0) | (full.SibSp > 0))].value_counts()
table_surname['Surname_surviving_men'] = table_surname['Surname_surviving_men'].fillna(0)
table_surname['Surname_surviving_men'][table_surname['Surname_surviving_men'] > 0] = 1.0
table_surname["Surname_Id"]= pd.Categorical.from_array(table_surname.index).codes
# compress under 3 members into one code.
table_surname["Surname_Id"][table_surname["Surname_Members"] < 3 ] = -1
table_surname["Surname_Members"] = pd.cut(table_surname["Surname_Members"], bins=[0,1,4,20], labels=[0,1,2])
full = pd.merge(full, table_surname, left_on="surname",right_index=True,how='left', sort=False)
### AGE PROCESSING --------------------------------------------------------------------------
classers = ['Fare','Parch','Pclass','SibSp','TitleCat',
'CabinCat','female','male', 'Embarked', 'FamilySize', 'NameLength','Ticket_Members','Ticket_Id']
etr = ExtraTreesRegressor(n_estimators=200)
X_train = full[classers][full['Age'].notnull()]
Y_train = full['Age'][full['Age'].notnull()]
X_test = full[classers][full['Age'].isnull()]
etr.fit(X_train,np.ravel(Y_train))
age_preds = etr.predict(X_test)
full['Age'][full['Age'].isnull()] = age_preds
# FEATURES -----------------------------------------------------------------------------------
features = ['female','male','Age','male_adult','female_adult', 'child','TitleCat', 'Pclass',
'Pclass','Ticket_Id','NameLength','CabinType','CabinCat', 'SibSp', 'Parch',
'Fare','Embarked','Surname_Members','Ticket_Members','FamilySize',
'Ticket_perishing_women','Ticket_surviving_men',
'Surname_perishing_women','Surname_surviving_men']
train = full[0:891].copy()
test = full[891:].copy()
selector = SelectKBest(f_classif, k=len(features))
selector.fit(train[features], target)
scores = -np.log10(selector.pvalues_)
indices = np.argsort(scores)[::-1]
print("Features importance :")
for f in range(len(scores)):
print("%0.2f %s" % (scores[indices[f]],features[indices[f]]))
# BEST CLASSIFIER METHOD ==> RANDOM FOREST -----------------------------------------------------
rfc = RandomForestClassifier(n_estimators=3000, min_samples_split=4, class_weight={0:0.745,1:0.255})
# CROSS VALIDATION WITH RANDOM FOREST CLASSIFIER METHOD-----------------------------------------
kf = model_selection.KFold(train.shape[0], n_folds=3, random_state=1)
scores = model_selection.cross_val_score(rfc, train[features], target, cv=kf)
print("Accuracy: %0.3f (+/- %0.2f) [%s]" % (scores.mean()*100, scores.std()*100, 'RFC Cross Validation'))
rfc.fit(train[features], target)
score = rfc.score(train[features], target)
print("Accuracy: %0.3f [%s]" % (score*100, 'RFC full test'))
importances = rfc.feature_importances_
indices = np.argsort(importances)[::-1]
for f in range(len(features)):
print("%d. feature %d (%f) %s" % (f + 1, indices[f]+1, importances[indices[f]]*100, features[indices[f]]))
# PREDICTION -----------------------------------------------------------------------------------
rfc.fit(train[features], target)
predictions = rfc.predict(test[features])
# OUTPUT FILE -----------------------------------------------------------------------------------
PassengerId =np.array(test["PassengerId"]).astype(int)
my_prediction = pd.DataFrame(predictions, PassengerId, columns = ["Survived"])
my_prediction.to_csv("my_prediction.csv", index_label = ["PassengerId"])
print("The end ...")