-
Notifications
You must be signed in to change notification settings - Fork 0
/
ohe_on_traintestdata.py
236 lines (159 loc) · 7.66 KB
/
ohe_on_traintestdata.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
import numpy as np
import pandas as pd
from sklearn.preprocessing import OneHotEncoder
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression
from sklearn.linear_model import LogisticRegressionCV
from sklearn.metrics import roc_auc_score
from sklearn.model_selection import StratifiedKFold
from sklearn.model_selection import GridSearchCV
import warnings
warnings.filterwarnings("ignore")
train = pd.read_csv("Kaggle/catinthedat/train.csv")
test = pd.read_csv("Kaggle/catinthedat/test.csv")
sample_sub_df = pd.read_csv("Kaggle/catinthedat/sample_submission.csv")
train_id = train["id"]
test_id = test["id"]
target = train["target"]
train.drop(["id","target","bin_0"], axis = 1,inplace = True)
test.drop(["id","bin_0"], axis = 1,inplace = True)
# nom_0
train = pd.get_dummies(train, columns = ["nom_0"])
test = pd.get_dummies(test, columns = ["nom_0"])
nom_0_Blue_dict = {0:0,1:255}
nom_0_Green_dict = {0:0,1:128}
nom_0_Red_dict = {0:0,1:255}
train["nom_0_Blue"] = train["nom_0_Blue"].map(nom_0_Blue_dict)
train["nom_0_Green"] = train["nom_0_Green"].map(nom_0_Green_dict)
train["nom_0_Red"] = train["nom_0_Red"].map(nom_0_Red_dict)
test["nom_0_Blue"] = test["nom_0_Blue"].map(nom_0_Blue_dict)
test["nom_0_Green"] = test["nom_0_Green"].map(nom_0_Green_dict)
test["nom_0_Red"] = test["nom_0_Red"].map(nom_0_Red_dict)
bin34_dict = {"T":1,"F":0,"Y":1,"N":0}
#bin_3
train["bin_3"] = train["bin_3"].map(bin34_dict)
test["bin_3"] = test["bin_3"].map(bin34_dict)
#bin_4
train["bin_4"] = train["bin_4"].map(bin34_dict)
test["bin_4"] = test["bin_4"].map(bin34_dict)
# ord_5 Method Trials:
## Method 1
train["ord_5_1"] = train["ord_5"].str[0]
train["ord_5_2"] = train["ord_5"].str[1]
test["ord_5_1"] = test["ord_5"].str[0]
test["ord_5_2"] = test["ord_5"].str[1]
train.drop(["ord_5","ord_5_2"],axis = 1, inplace = True)
test.drop(["ord_5","ord_5_2"],axis = 1, inplace = True)
# Ascii Application
import string
for column in ["ord_3","ord_4","ord_5_1"]:
train[column] = train[column].apply(lambda x: string.ascii_letters.index(x))
test[column] = test[column].apply(lambda x: string.ascii_letters.index(x))
## Method 2
#ord_5 = sorted(list(set(train["ord_5"].values)))
#ord_5 = dict(zip(ord_5,range(len(ord_5))))
#train["ord_5"] = train["ord_5"].apply(lambda x: ord_5[x]).astype(int)
#test["ord_5"] = test["ord_5"].apply(lambda x: ord_5[x]).astype(int)
#ord_3 = sorted(list(set(train["ord_3"].values)))
#ord_3 = dict(zip(ord_3,range(len(ord_3))))
#train["ord_3"] = train["ord_3"].apply(lambda x: ord_3[x]).astype(int)
#test["ord_3"] = test["ord_3"].apply(lambda x: ord_3[x]).astype(int)
#ord_4 = sorted(list(set(train["ord_4"].values)))
#ord_4 = dict(zip(ord_4,range(len(ord_4))))
#train["ord_4"] = train["ord_4"].apply(lambda x: ord_4[x]).astype(int)
#test["ord_4"] = test["ord_4"].apply(lambda x: ord_4[x]).astype(int)
## Method 3
#from sklearn.preprocessing import OrdinalEncoder # Encode categorical features as an integer array
#ordinal_columns = ["ord_5","ord_3","ord_4"]
#orden = OrdinalEncoder(categories = "auto")
#orden.fit(train[ordinal_columns])
#train[ordinal_columns] = orden.transform(train[ordinal_columns])
#test[ordinal_columns] = orden.transform(test[ordinal_columns])
# Replace values that are both not in train and test
columns_to_check = ["nom_7","nom_8","nom_9"]
#
for column in columns_to_check:
values_to_replace = set(train[column]) ^ set(test[column])
if values_to_replace:
print("Column " + column + " has " + str(len(values_to_replace)) + " values")
train[column] = train[column].apply(lambda x: "21578c358" if x in values_to_replace else x)
test[column] = test[column].apply(lambda x: "21578c358" if x in values_to_replace else x)
else:
print("Column" + column + "has no none existent values")
# ord_1
ord1_dict = {"Novice":1,"Contributor":2,"Expert":3,"Master":4,"Grandmaster":5}
train["ord_1"] = train["ord_1"].map(ord1_dict)
ord1_dict = {"Novice":1,"Contributor":2,"Expert":3,"Master":4,"Grandmaster":5}
test["ord_1"] = test["ord_1"].map(ord1_dict)
#ord_2
ord2_dict = {"Freezing":1,"Cold":2,"Warm":3,"Hot":4,"Boiling Hot":5,"Lava Hot":6}
train["ord_2"] = train["ord_2"].map(ord2_dict)
ord2_dict = {"Freezing":1,"Cold":2,"Warm":3,"Hot":4,"Boiling Hot":5,"Lava Hot":6}
test["ord_2"] = test["ord_2"].map(ord2_dict)
#Month & Day
# Month sin-cosin transformation
train['mnth_sin'] = np.sin((train.month-2)*(2.*np.pi/12.0))
train['mnth_cos'] = np.cos((train.month-2)*(2.*np.pi/12.0))
test['mnth_sin'] = np.sin((test.month-2)*(2.*np.pi/12.0))
test['mnth_cos'] = np.cos((test.month-2)*(2.*np.pi/12.0))
# Day sin-cosin transformation
train['day_sin'] = np.sin((train.day-2)*(2.*np.pi/7.0))
train['day_cos'] = np.cos((train.day-2)*(2.*np.pi/7.0))
test['day_sin'] = np.sin((test.day-2)*(2.*np.pi/7.0))
test['day_cos'] = np.cos((test.day-2)*(2.*np.pi/7.0))
train.drop(["month","day"], axis = 1,inplace = True)
test.drop(["month","day"], axis = 1,inplace = True)
# Convert nom_5 - nom_9 into integers
hexadecimal_columns = ["nom_5","nom_6","nom_7","nom_8","nom_9"]
import binascii
for column in hexadecimal_columns:
train[column] = train[column].apply(lambda x: int(x,36))
test[column] = test[column].apply(lambda x: int(x,36))
# Standard Scaler
from sklearn.preprocessing import StandardScaler
train_to_be_scaled = train[['mnth_sin', 'mnth_cos', 'day_sin', 'day_cos']]
test_to_be_scaled = test[['mnth_sin', 'mnth_cos', 'day_sin', 'day_cos']]
se = StandardScaler()
se.fit(train_to_be_scaled)
train_to_be_scaled = se.transform(train_to_be_scaled)
test_to_be_scaled = se.transform(test_to_be_scaled)
train_to_be_scaled_df = pd.DataFrame(train_to_be_scaled,columns = ['mnth_sin', 'mnth_cos', 'day_sin', 'day_cos'])
test_to_be_scaled_df = pd.DataFrame(test_to_be_scaled, columns = ['mnth_sin', 'mnth_cos', 'day_sin', 'day_cos'])
train = pd.concat([train.drop(['mnth_sin', 'mnth_cos', 'day_sin', 'day_cos'],axis = 1),train_to_be_scaled_df],axis = 1)
test = pd.concat([test.drop(['mnth_sin', 'mnth_cos', 'day_sin', 'day_cos'],axis = 1),test_to_be_scaled_df],axis = 1)
#dummy encoding
traintest = pd.concat([train,test])
dummies = pd.get_dummies(traintest,columns = traintest.columns, drop_first = True,sparse = True)
train_ohe = dummies.iloc[:train.shape[0],:]
test_ohe = dummies.iloc[train.shape[0]:,:]
train_ohe = train_ohe.sparse.to_coo().tocsr()
test_ohe = test_ohe.sparse.to_coo().tocsr()
# Submission
%%time
lr = LogisticRegression(C=0.123456789,
solver="lbfgs",
max_iter=70000,
random_state = 42,
intercept_scaling = 0.1,
tol=0.0002,
class_weight = "balanced")
lr.fit(train_ohe,target)
sample_sub_df['target'] = lr.predict_proba(test_ohe)[:,1]
sample_sub_df.to_csv('subm102.csv', index=False)
# StratifiedKFold cv score
folds = StratifiedKFold(n_splits = 5,shuffle = True,random_state = 42)
model = LogisticRegression(C=0.123456789,#C=0.12356789
solver="lbfgs",
max_iter=70000,
random_state = 42,
intercept_scaling =0.1,
tol=0.0002)
final_scores = []
for fold, (train_index,validation_index) in enumerate(folds.split(train_ohe,target,groups = target)):
print("Fold:", fold+1)
train_x,train_y = train_ohe[train_index,:], target[train_index]
valid_x,valid_y = train_ohe[validation_index,:], target[validation_index]
%time model.fit(train_x,train_y)
predictions = model.predict(valid_x)
final_scores.append(roc_auc_score(valid_y,predictions))
print("AUC SCORE: {}".format(roc_auc_score(valid_y,predictions)))