-
Notifications
You must be signed in to change notification settings - Fork 0
/
model.py
executable file
·211 lines (145 loc) · 5.07 KB
/
model.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
#!/usr/bin/env python3
import matplotlib.pyplot as plt
import missingno as msno
import sys
import numpy as np
import sklearn as sk
from sklearn.ensemble import RandomForestClassifier
import pandas as pd
TEST_DATA = 'data/test.csv'
TRAIN_DATA = 'data/train.csv'
nan_count = 0
def main():
train_data = pd.read_csv(TRAIN_DATA)
train(train_data)
def train(train_data):
train_X = clean_data(train_data.drop('Survived', axis=1))
train_Y = train_data['Survived']
model, accuracy = train_model(train_X, train_Y)
print(accuracy)
if '--test' in sys.argv:
file_path = sys.argv[2]
print(f'writing answers to {file_path}')
test_model(model, file_path)
# returns A - H
def a_to_h():
return list(map(lambda n: chr(97 + n).upper(), range(0, 8)))
def non_numeric_ticket_blobs(X):
blobs = set()
nondistinct_blobs = list(X.apply(grab_non_numeric_blobs, axis=1))
for blob in nondistinct_blobs:
for word in blob:
blobs.add(word)
return blobs
def grab_non_numeric_blobs(row):
blob_set = set()
ticket = row['Ticket']
blobs = ticket.split(' ')
for blob in blobs:
if not blob.isnumeric():
blob_set.add(blob)
return blob_set
def test_model(model, results_file):
test_X = clean_data(pd.read_csv(TEST_DATA))
test_X.to_csv('test_x.csv')
pred_Y = model.predict(test_X)
test_X['Survived'] = pred_Y
test_X[['Survived', 'PassengerId']].to_csv(results_file, index=False)
def clean_data(X):
msno.matrix(X)
plt.show()
ticket_columns = set(build_ticket_columns())
print(ticket_columns)
ticket_values = X[['Ticket']]
ticket_values['Ticket'].transform(lambda e: e.replace('STON', 'SOTON'))
train_X = X.drop('Name', axis=1).drop('Ticket', axis=1)
columns = train_X.columns
mean_age = train_X['Age'].median()
mean_fare = train_X['Fare'].median()
ticket_column_values = ticket_values\
.apply(lambda r: derive_ticket_columns(r, ticket_columns), axis=1)\
.drop('Ticket', axis=1)
train_X = train_X.apply(lambda r: set_age_as_mean(r, mean_age), axis=1)\
.apply(lambda r: set_fare_as_mean(r, mean_fare), axis=1)\
.join(create_cabin_columns(X)).drop('Cabin', axis=1)\
.join(ticket_column_values)
train_X = pd.get_dummies(train_X)
train_X.to_csv('train_x.csv')
return train_X
def derive_ticket_columns(row, column_names):
numbers = list(filter(lambda s: s.isnumeric() and len(s) > 3, str(row['Ticket']).split(' ')))
if len(numbers) == 0:
row['TicketNumber'] = -1
else:
row['TicketNumber'] = float(numbers[0])
for name in column_names:
if name in row['Ticket']:
row[str(name)] = 1
else:
row[str(name)] = 0
return row
# key: 'Cabin'
def create_cabin_columns(df):
dictionaries = df.apply(create_cabin_column_dict, axis=1)
result_dict = dict()
for dictionary in dictionaries:
for key, value in dictionary.items():
if key in result_dict:
result_dict[key].append(value)
else:
result_dict[key] = [value]
return pd.DataFrame.from_dict(result_dict)
def cabin_number(row):
segments = str(row['Cabin']).split(' ')
numeric_segments = []
for segment in segments:
numeric_segment = ''
for c in segment:
if c.isnumeric():
numeric_segment += c
if len(numeric_segment) != 0:
numeric_segments.append(numeric_segment)
if len(numeric_segments) != 0:
return float(numeric_segments[0])
else:
return -1
def create_cabin_column_dict(row):
letters = a_to_h()
cabin_column_values = dict()
number = cabin_number(row)
cabin_column_values['CabinNumber'] = number
for letter in letters:
if letter in str(row['Cabin']):
cabin_column_values[f'Cabin_{letter}'] = 1
else:
cabin_column_values[f'Cabin_{letter}'] = 0
return cabin_column_values
def train_model(train_X, train_Y):
model = RandomForestClassifier(n_estimators=512)
model.fit(train_X, train_Y)
y_predictions = model.predict(train_X)
accuracy = sk.metrics.accuracy_score(train_Y, y_predictions)
return (model, accuracy)
def build_ticket_columns():
test = pd.read_csv(TEST_DATA)[['Ticket']]
train = pd.read_csv(TRAIN_DATA)[['Ticket']]
data = pd.concat((test, train), axis=1)
def extract_destinations(row):
ticket_text = row['Ticket']
if not pd.isna(ticket_text[0]):
chunks = str(ticket_text[0]).split(' ')
else:
chunks = []
return list(filter(lambda s: not s.isnumeric(), chunks))
chunk_list = list(data.apply(extract_destinations, axis=1))
return np.concatenate(chunk_list)
def set_age_as_mean(row, mean):
if pd.isna(row['Age']):
row['Age'] = mean
return row
def set_fare_as_mean(row, mean):
if pd.isna(row['Fare']):
row['Fare'] = mean
return row
if __name__ == '__main__':
main()