-
Notifications
You must be signed in to change notification settings - Fork 0
/
Titanic_boost_test4.py
274 lines (199 loc) · 10.2 KB
/
Titanic_boost_test4.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
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
# -*- coding: utf-8 -*-
"""
Created on Fri Apr 18 11:29:36 2020
@author: damar
"""
import numpy as np
import pandas as pd
import pprint
import matplotlib.pyplot as plt
import re
search_type = "video"
#for year in Years:
#datasetpivot = dataset.stack()
#Load data if no header
#header = ['Country','Year','Value', 'Value Footnotes']
#Populations = pd.read_csv(Path+r'\UNData_Population.csv', sep=',', header=1, names=header,engine='python')
# No file path load
#Titanic_Train = pd.read_csv('train.csv', sep=',',engine='python')
# file path load
Path=r"C:\Users\damar\Documents\Python Scripts\Titanic"
Titanic_Train = pd.read_csv(Path+r'\train.csv', sep=',',engine='python')
Titanic_Test = pd.read_csv(Path+r'\test.csv', sep=',',engine='python')
Titanic_Test.head()
#Reorder class to 1st class better than 3
#Mapper = {1:3, 3:1}
#Titanic_Train['Pclass'] = Titanic_Train['Pclass'].replace(Mapper)
#Titanic_Test['Pclass'] = Titanic_Test['Pclass'].replace(Mapper)
#Is Married woman indicator
Titanic_Train['Is_Married_Fem'] = [ 1 if 'Mrs' in x else 0 for x in Titanic_Train['Name']]
Titanic_Test['Is_Married_Fem'] = [ 1 if 'Mrs' in x else 0 for x in Titanic_Test['Name']]
#Is First Class
Titanic_Train['Is_Firstclass'] = [ 1 if x == 1 else 0 for x in Titanic_Train['Pclass']]
Titanic_Test['Is_Firstclass'] = [ 1 if x == 1 else 0 for x in Titanic_Test['Pclass']]
#Convert <1 ages
Titanic_Train['Is_kid'] = [ 1 if 'Master' in x else 0 for x in Titanic_Train['Name']]
Titanic_Test['Is_kid'] = [ 1 if 'Master' in x else 0 for x in Titanic_Test['Name']]
#Has >3 siblings spouse
Titanic_Train['Has_Siblings'] = [ 1 if x > 1 else 0 for x in Titanic_Train['SibSp']]
Titanic_Test['Has_Siblings'] = [ 1 if x > 1 else 0 for x in Titanic_Test['SibSp']]
#Has >3 Has Parent children
Titanic_Train['Has_Parent'] = [ 1 if x > 1 else 0 for x in Titanic_Train['Parch']]
Titanic_Test['Has_Parent'] = [ 1 if x > 1 else 0 for x in Titanic_Test['Parch']]
Titanic_Train['Has_Family'] = Titanic_Train['SibSp'] + Titanic_Train['Parch']
Titanic_Test['Has_Family'] = Titanic_Test['SibSp'] + Titanic_Test['Parch']
#Titanic_Train['Is_Elderly'] = [ 1 if x > 60 else 0 for x in Titanic_Train['Age']]
#Titanic_Test['Is_Elderly'] = [ 1 if x > 60 else 0 for x in Titanic_Test['Age']]
Mapper = {np.nan:5}
#Titanic_Train['Age'] = Titanic_Train[Titanic_Train['Is_kid'] == 1]['Age'].replace(Mapper)
#Titanic_Test['Age'] = Titanic_Test[Titanic_Test['Is_kid'] == 1]['Age'].replace(Mapper)
#Titanic_Train['Age'] = [ 5 if x<1 else x for x in Titanic_Train['Age']]
#Titanic_Test['Age'] = [ 5 if x<1 else x for x in Titanic_Test['Age']]
#Titanic_Train[Titanic_Train['Is_kid']==1] = Titanic_Train[Titanic_Train['Is_kid']==1].replace(np.nan, 9)
#Titanic_Test[Titanic_Test['Is_kid']==1] = Titanic_Test[Titanic_Test['Is_kid']==1].replace(np.nan, 9)
#Titanic_Train[Titanic_Train['PassengerId'] == 79]['Age'].head()
#Titanic_Train['Age'] = [ 3 if x<1 else x for x in Titanic_Train['Age']]
#Titanic_Test['Age'] = [ 3 if x<1 else x for x in Titanic_Test['Age']]
Titanic_Train.info()
#X = Titanic_Train[['Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare', 'Embarked','Is_Married_Fem', 'Is_Firstclass', 'Is_kid']]
#X_train = Titanic_Train.iloc[:, [2,4,5,9,11,12,13,14,15,16]].values
#y_train = Titanic_Train.iloc[:,1].values
#X_train = Titanic_Train.loc[:, ['Pclass', 'Sex','Age','Fare','Embarked','Is_Married_Fem','Is_kid','Has_Siblings','Has_Parent']]
X_train = Titanic_Train.loc[:, ['Pclass','Name', 'Sex','Age','Fare','Embarked','Is_Married_Fem', 'Is_kid', 'Has_Family']]
y_train = Titanic_Train.loc[:,['Survived']]
#X_test = Titanic_Test.loc[:, ['Pclass', 'Sex','Age','Fare','Embarked','Is_Married_Fem','Is_kid','Has_Siblings','Has_Parent']]
X_test = Titanic_Test.loc[:, ['Pclass','Name', 'Sex','Age','Fare','Embarked','Is_Married_Fem', 'Is_kid', 'Has_Family']]
#X_test = Titanic_Test.iloc[:, [1,3,4,8,10,11,12,13, 14,15]].values
#print(X_test)
#y_test = Titanic_Test.iloc[:,1].values
median_list = list(X_train.groupby('Pclass')['Age'].median().values)
for i in range(3):
X_train.loc[X_train['Pclass']==i+1,'Age'] = X_train.loc[X_train['Pclass']==i+1,'Age'].fillna(median_list[i])
#median_list = list(X_test.groupby('Pclass')['Age'].median().values)
for i in range(3):
X_test.loc[X_test['Pclass']==i+1,'Age'] = X_test.loc[X_test['Pclass']==i+1,'Age'].fillna(median_list[i])
X_train['age_range'] = pd.cut(X_train['Age'], bins=[0,2,17,55,99], labels=['Baby', 'Child', 'Adult', 'Elderly'])
X_test['age_range'] = pd.cut(X_test['Age'], bins=[0,2,17,55,99], labels=['Baby', 'Child', 'Adult', 'Elderly'])
#Impute Null age using Mean strategy
from sklearn.impute import SimpleImputer
imputer = SimpleImputer(missing_values=np.nan, strategy = 'median')
#imputer.fit(X_train.loc[:,['Age']])
#X_train.loc[:,['Age']] = imputer.transform(X_train.loc[:,['Age']])
#X_test.loc[:,['Age']] = imputer.transform(X_test.loc[:,['Age']])
#Impute Null fare using Mean strategy
imputer.fit(X_train.loc[:,['Fare']])
X_train.loc[:,['Fare']] = imputer.transform(X_train.loc[:,['Fare']])
X_test.loc[:,['Fare']] = imputer.transform(X_test.loc[:,['Fare']])
#Impute Null embark using most_frequent strategy
imputer = SimpleImputer(missing_values=np.nan, strategy = 'most_frequent')
imputer.fit(X_train.loc[:,['Embarked']])
X_train.loc[:,['Embarked']] = imputer.transform(X_train.loc[:,['Embarked']])
X_test.loc[:,['Embarked']] = imputer.transform(X_test.loc[:,['Embarked']])
#Encode Sex Label
from sklearn.preprocessing import LabelEncoder
le_1 = LabelEncoder()
X_train.loc[:, ['Sex']] = le_1.fit_transform(X_train.loc[:, ['Sex']])
X_test.loc[:, ['Sex']] = le_1.transform(X_test.loc[:, ['Sex']])
title_mapping = {'Don':'Rare','Rev':'Rare','Mme':'Miss','Ms':'Miss',
'Major':'Rare','Lady':'Royal','Mlle':'Miss','Col':'Rare','Capt':'Rare',
'Sir':'Royal', 'Countess':'Royal', 'Jonkheer':'Royal', 'Dona':'Royal'}
for i in X_train['Name']:
X_train['Title'] = X_train['Name'].str.extract('([A-Za-z]+)\.',expand=True)
# Dropping Name
X_train.drop(columns=['Name'],inplace=True)
# Replacing by mapping
X_train.groupby('Title').count()
X_train.replace({'Title':title_mapping},inplace=True)
for i in X_test['Name']:
X_test['Title'] = X_test['Name'].str.extract('([A-Za-z]+)\.',expand=True)
# Dropping Name
X_test.drop(columns=['Name'],inplace=True)
# Replacing by mapping
#X_test.groupby('Title').count()
X_test.replace({'Title':title_mapping},inplace=True)
#print(X_test.Title.unique())
#Onehotencoder Pclass field 1, 2, 3
dummies = pd.get_dummies(X_train['Pclass'], prefix='Pclass', drop_first= True)
X_train= pd.concat([X_train, dummies], axis=1)
X_train = X_train.drop('Pclass', axis=1)
dummies = pd.get_dummies(X_test['Pclass'], prefix='Pclass', drop_first= True)
X_test= pd.concat([X_test, dummies], axis=1)
X_test = X_test.drop('Pclass', axis=1)
#Onehotencoder Title Mr, Mrs, Miss, Master, Rare, Dr
dummies = pd.get_dummies(X_train['Title'], prefix='Title', drop_first= True)
X_train= pd.concat([X_train, dummies], axis=1)
X_train = X_train.drop('Title', axis=1)
dummies = pd.get_dummies(X_test['Title'], prefix='Title', drop_first= True)
X_test= pd.concat([X_test, dummies], axis=1)
X_test = X_test.drop('Title', axis=1)
dummies = pd.get_dummies(X_train['age_range'], prefix='age_range', drop_first= True)
X_train= pd.concat([X_train, dummies], axis=1)
X_train = X_train.drop('age_range', axis=1)
dummies = pd.get_dummies(X_test['age_range'], prefix='age_range', drop_first= True)
X_test= pd.concat([X_test, dummies], axis=1)
X_test = X_test.drop('age_range', axis=1)
#Onehotencoder Embark field S, C, Q
dummies = pd.get_dummies(X_train['Embarked'], prefix='Embarked', drop_first= True)
X_train= pd.concat([X_train, dummies], axis=1)
#Removed column already encoded
X_train = X_train.drop('Embarked', axis=1)
dummies = pd.get_dummies(X_test['Embarked'], prefix='Embarked', drop_first= True)
X_test= pd.concat([X_test, dummies], axis=1)
#Removed column already encoded
X_test = X_test.drop('Embarked', axis=1)
#Train test split
#from sklearn.model_selection import train_test_split
#X_train, X_test, y_train, y_test = train_test_split(X, y,test_size=0.2, random_state=0)
#Apply scaling using Standardization
from sklearn.preprocessing import StandardScaler
sc = StandardScaler()
#X = sc.fit_transform(X)
X_train = sc.fit_transform(X_train)
X_test = sc.transform(X_test)
import xgboost as xgb
D_train = xgb.DMatrix(X_train, label=y_train)
D_test = xgb.DMatrix(X_test)
"""
param = {
'eta': 0.3,
'max_depth': 15,
'gamma': 0.4,
'min_child_weight': 5,
'objective': 'multi:softmax',
'num_class': 2}
"""
param = {
'eta': 0.15,
'max_depth': 10,
'gamma': 0.2,
'min_child_weight': 5,
"objective": 'multi:softmax',
'num_class': 2,
'n_jobs':0,
'random_state':0,
"colsample_bytree":0.5
}
steps = 30 # The number of training iterations
#model = xgb.train(param, D_train, steps)
"""
model = xgb.XGBClassifier(objective='multi:softmax', eta=0.15,
gamma=0.2,random_state=0, n_estimators=100, max_depth=15,
min_child_weight=5, max_features=0.7,
num_class =2,
min_samples_leaf=0.6, importance_type='gain',
learning_rate=0.5, colsample_bytree=0.6)
"""
model = xgb.XGBClassifier(objective='reg:logistic', eta=0.05,
gamma=0.4,random_state=0, n_estimators=100, max_depth=5,
min_child_weight=1, max_features=0.4,
base_score = 0.5,
min_samples_leaf=0.6, importance_type='gain',
learning_rate=0.1, colsample_bytree=0.3)
model.fit(X_train, y_train)
#Predict test set result
y_pred = model.predict(X_test)
#y_pred = np.asarray([np.argmax(line) for line in D_pred])
header = ['Survived']
Titanic_Pred = pd.concat([Titanic_Test,pd.DataFrame(y_pred, columns=header)],axis=1)
Titanic_csvfile = Path+'\Titanic_submission41.csv'
Titanic_Pred[['PassengerId', 'Survived']].to_csv(Titanic_csvfile, mode='w', index=False)