-
Notifications
You must be signed in to change notification settings - Fork 91
/
all_to7.py
328 lines (281 loc) · 19.2 KB
/
all_to7.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
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
# coding: UTF-8
import pandas as pd
import time
from xgboost import XGBClassifier
import xgboost as xgb
import numpy as np
from datetime import datetime,timedelta
import pickle,os
from dateutil.parser import parse
from sklearn.utils import shuffle
from sklearn.metrics import roc_auc_score
import lightgbm as lgb
import datetime
from sklearn.linear_model import LogisticRegression
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
data_train=pd.read_csv('../data/round2_ijcai_18_train_20180425.txt',delim_whitespace=True)
#data_test_a=pd.read_csv('/test_a_trade_250.txt',delim_whitespace=True)
#data_test_a['is_trade'] = 0
#data_train=pd.concat([data_train,data_test_a]).reset_index(drop=True)
#data_train=data_train.drop_duplicates(['instance_id'])
data_train['item_category_1'],data_train['item_category_2'],data_train['item_category_3']=data_train['item_category_list'].str.split(';',2).str
del data_train['item_category_list']
data_train['predict_category_property_A'],data_train['predict_category_property_B'],data_train['predict_category_property_C']=data_train['predict_category_property'].str.split(';',2).str
del data_train['predict_category_property']
data_train['predict_category_property_A'],data_train['predict_category_property_A_1']=data_train['predict_category_property_A'].str.split(':',1).str
data_train['predict_category_property_A_1'],data_train['predict_category_property_A_2'],data_train['predict_category_property_A_3']=data_train['predict_category_property_A_1'].str.split(',',2).str
del data_train['predict_category_property_A_3']
data_train['predict_category_property_B'],data_train['predict_category_property_B_1']=data_train['predict_category_property_B'].str.split(':',1).str
data_train['predict_category_property_B_1'],data_train['predict_category_property_B_2'],data_train['predict_category_property_B_3']=data_train['predict_category_property_B_1'].str.split(',',2).str
del data_train['predict_category_property_B_3']
data_train['predict_category_property_C'],data_train['predict_category_property_C_1']=data_train['predict_category_property_C'].str.split(':',1).str
data_train['predict_category_property_C_1'],data_train['predict_category_property_C_2'],data_train['predict_category_property_C_3']=data_train['predict_category_property_C_1'].str.split(',',2).str
data_train['predict_category_property_C_1'],data_train['predict_category_property_C_3']=data_train['predict_category_property_C_1'].str.split(';',1).str
data_train['predict_category_property_C_2'],data_train['predict_category_property_C_3']=data_train['predict_category_property_C_2'].str.split(';',1).str
del data_train['predict_category_property_C_3']
#del data_train['predict_category_property_C']
data_train['item_property_list_1'],data_train['item_property_list_2'],data_train['item_property_list_3'],data_train['item_property_list_4']=data_train['item_property_list'].str.split(';',3).str
del data_train['item_property_list_4']
del data_train['item_property_list']
del data_train['item_category_3']
#data_train=data_train.fillna(-1)
##处理类目类特征，将类目类特征分为一列一列
#data_test=pd.read_csv('../data/round2_ijcai_18_test_a_20180425.txt',delim_whitespace=True)
print('load data_test...')
data_test_b=pd.read_csv('../data/round2_ijcai_18_test_b_20180510.txt',delim_whitespace=True)
print('data_test_b:%d' %(len(data_test_b)))
data_test_a=pd.read_csv('../data/round2_ijcai_18_test_a_20180425.txt',delim_whitespace=True)
print('data_test_a:%d' %(len(data_test_a)))
data_test=pd.concat([data_test_b,data_test_a]).reset_index(drop=True)
data_test['item_category_1'],data_test['item_category_2'],data_test['item_category_3']=data_test['item_category_list'].str.split(';',2).str
del data_test['item_category_list']
data_test['predict_category_property_A'],data_test['predict_category_property_B'],data_test['predict_category_property_C']=data_test['predict_category_property'].str.split(';',2).str
del data_test['predict_category_property']
data_test['predict_category_property_A'],data_test['predict_category_property_A_1']=data_test['predict_category_property_A'].str.split(':',1).str
data_test['predict_category_property_A_1'],data_test['predict_category_property_A_2'],data_test['predict_category_property_A_3']=data_test['predict_category_property_A_1'].str.split(',',2).str
del data_test['predict_category_property_A_3']
data_test['predict_category_property_B'],data_test['predict_category_property_B_1']=data_test['predict_category_property_B'].str.split(':',1).str
data_test['predict_category_property_B_1'],data_test['predict_category_property_B_2'],data_test['predict_category_property_B_3']=data_test['predict_category_property_B_1'].str.split(',',2).str
del data_test['predict_category_property_B_3']
data_test['predict_category_property_C'],data_test['predict_category_property_C_1']=data_test['predict_category_property_C'].str.split(':',1).str
data_test['predict_category_property_C_1'],data_test['predict_category_property_C_2'],data_test['predict_category_property_C_3']=data_test['predict_category_property_C_1'].str.split(',',2).str
data_test['predict_category_property_C_1'],data_test['predict_category_property_C_3']=data_test['predict_category_property_C_1'].str.split(';',1).str
data_test['predict_category_property_C_2'],data_test['predict_category_property_C_3']=data_test['predict_category_property_C_2'].str.split(';',1).str
del data_test['predict_category_property_C_3']
#del data_test['predict_category_property_C']
data_test['item_property_list_1'],data_test['item_property_list_2'],data_test['item_property_list_3'],data_test['item_property_list_4']=data_test['item_property_list'].str.split(';',3).str
del data_test['item_property_list_4']
del data_test['item_property_list']
data_train.to_csv('../data/data_train_ori.csv')
data_test.to_csv('../data/data_test_ori.csv')
print('load_data_ori...')
data_train=pd.read_csv('../data/data_train_ori_b.csv')
print(len(data_train))
#data_train['times']=data_train['times'].astype(str)
data_test=pd.read_csv('../data/data_test_ori_b.csv')
print(len(data_test))
#data_test['times']=data_test['times'].astype(str)
def convert_data(data):
data["times"] = data["context_timestamp"].apply(lambda x: datetime.datetime.fromtimestamp(x))
data["day"] = data["times"].apply(lambda x: x.day)
data["hour"] = data["times"].apply(lambda x: x.hour)
data['min'] = data['times'].apply(lambda x: x.minute)
data['day']=data['day'].astype('int')
data['hour']=data['hour'].astype('int')
data['min']=data['min'].astype('int')
# 小时均值特征
grouped = data.groupby('user_id')['hour'].mean().reset_index()
grouped.columns = ['user_id', 'user_mean_hour']
data = data.merge(grouped, how='left', on='user_id')
grouped = data.groupby('item_id')['hour'].mean().reset_index()
grouped.columns = ['item_id', 'item_mean_hour']
data = data.merge(grouped, how='left', on='item_id')
grouped = data.groupby('shop_id')['hour'].mean().reset_index()
grouped.columns = ['shop_id', 'shop_mean_hour']
data = data.merge(grouped, how='left', on='shop_id')
grouped = data.groupby('item_city_id')['hour'].mean().reset_index()
grouped.columns = ['item_city_id', 'city_mean_hour']
data = data.merge(grouped, how='left', on='item_city_id')
grouped = data.groupby('item_brand_id')['hour'].mean().reset_index()
grouped.columns = ['item_brand_id', 'brand_mean_hour']
data = data.merge(grouped, how='left', on='item_brand_id')
# 小时var特征
'''
grouped = data.groupby('user_id')['hour'].var().reset_index()
grouped.columns = ['user_id', 'user_var_hour']
data = data.merge(grouped, how='left', on='user_id')
grouped = data.groupby('item_id')['hour'].var().reset_index()
grouped.columns = ['item_id', 'item_var_hour']
data = data.merge(grouped, how='left', on='item_id')
grouped = data.groupby('shop_id')['hour'].var().reset_index()
grouped.columns = ['shop_id', 'shop_var_hour']
data = data.merge(grouped, how='left', on='shop_id')
grouped = data.groupby('item_city_id')['hour'].var().reset_index()
grouped.columns = ['item_city_id', 'city_var_hour']
data = data.merge(grouped, how='left', on='item_city_id')
grouped = data.groupby('item_brand_id')['hour'].var().reset_index()
grouped.columns = ['item_brand_id', 'brand_var_hour']
data = data.merge(grouped, how='left', on='item_brand_id')
'''
#天均值特征
grouped = data.groupby('user_id')['day'].mean().reset_index()
grouped.columns = ['user_id', 'user_mean_day']
data = data.merge(grouped, how='left', on='user_id')
grouped = data.groupby('item_id')['day'].mean().reset_index()
grouped.columns = ['item_id', 'item_mean_day']
data = data.merge(grouped, how='left', on='item_id')
grouped = data.groupby('shop_id')['day'].mean().reset_index()
grouped.columns = ['shop_id', 'shop_mean_day']
data = data.merge(grouped, how='left', on='shop_id')
grouped = data.groupby('item_city_id')['day'].mean().reset_index()
grouped.columns = ['item_city_id', 'city_mean_day']
data = data.merge(grouped, how='left', on='item_city_id')
grouped = data.groupby('item_brand_id')['day'].mean().reset_index()
grouped.columns = ['item_brand_id', 'brand_mean_day']
data = data.merge(grouped, how='left', on='item_brand_id')
#天var特征
grouped = data.groupby('user_id')['day'].var().reset_index()
grouped.columns = ['user_id', 'user_var_day']
data = data.merge(grouped, how='left', on='user_id')
grouped = data.groupby('item_id')['day'].var().reset_index()
grouped.columns = ['item_id', 'item_var_day']
data = data.merge(grouped, how='left', on='item_id')
grouped = data.groupby('shop_id')['day'].var().reset_index()
grouped.columns = ['shop_id', 'shop_var_day']
data = data.merge(grouped, how='left', on='shop_id')
grouped = data.groupby('item_city_id')['day'].var().reset_index()
grouped.columns = ['item_city_id', 'city_var_day']
data = data.merge(grouped, how='left', on='item_city_id')
grouped = data.groupby('item_brand_id')['day'].var().reset_index()
grouped.columns = ['item_brand_id', 'brand_var_day']
data = data.merge(grouped, how='left', on='item_brand_id')
#天小时均值特征
grouped = data.groupby(['user_id', 'day'])['hour'].mean().reset_index()
grouped.columns = ['user_id', 'day', 'user_mean_day_hour']
data = data.merge(grouped, how='left', on=['user_id', 'day'])
grouped = data.groupby(['item_id', 'day'])['hour'].mean().reset_index()
grouped.columns = ['item_id', 'day', 'item_mean_day_hour']
data = data.merge(grouped, how='left', on=['item_id', 'day'])
grouped = data.groupby(['shop_id', 'day'])['hour'].mean().reset_index()
grouped.columns = ['shop_id', 'day', 'shop_mean_day_hour']
data = data.merge(grouped, how='left', on=['shop_id', 'day'])
grouped = data.groupby(['item_city_id', 'day'])['hour'].mean().reset_index()
grouped.columns = ['item_city_id', 'day', 'city_mean_day_hour']
data = data.merge(grouped, how='left', on=['item_city_id', 'day'])
grouped = data.groupby(['item_brand_id', 'day'])['hour'].mean().reset_index()
grouped.columns = ['item_brand_id', 'day', 'brand_mean_day_hour']
data = data.merge(grouped, how='left', on=['item_brand_id', 'day'])
#天小时var特征
grouped = data.groupby(['user_id', 'day'])['hour'].var().reset_index()
grouped.columns = ['user_id', 'day', 'user_var_day_hour']
data = data.merge(grouped, how='left', on=['user_id', 'day'])
grouped = data.groupby(['item_id', 'day'])['hour'].var().reset_index()
grouped.columns = ['item_id', 'day', 'item_var_day_hour']
data = data.merge(grouped, how='left', on=['item_id', 'day'])
grouped = data.groupby(['shop_id', 'day'])['hour'].var().reset_index()
grouped.columns = ['shop_id', 'day', 'shop_var_day_hour']
data = data.merge(grouped, how='left', on=['shop_id', 'day'])
grouped = data.groupby(['item_city_id', 'day'])['hour'].var().reset_index()
grouped.columns = ['item_city_id', 'day', 'city_var_day_hour']
data = data.merge(grouped, how='left', on=['item_city_id', 'day'])
grouped = data.groupby(['item_brand_id', 'day'])['hour'].var().reset_index()
grouped.columns = ['item_brand_id', 'day', 'brand_var_day_hour']
data = data.merge(grouped, how='left', on=['item_brand_id', 'day'])
return data
data_train["times"] = data_train["context_timestamp"].apply(lambda x: datetime.datetime.fromtimestamp(x))
data_train["day"] = data_train["times"].apply(lambda x: x.day)
data_test['is_trade'] = 0
print(len(data_test))
train_data = pd.concat([data_train,data_test],).reset_index(drop=True)
del data_train
del data_test
train_data["datetime"] = train_data["context_timestamp"].apply(lambda x: datetime.datetime.fromtimestamp(x))
train_data["day"] = train_data["datetime"].apply(lambda x: x.day)
train_data["hour"] = train_data["datetime"].apply(lambda x: x.hour)
##按时间排序
train_data = train_data.sort_values('context_timestamp')
train_data["item_category_list"] = train_data["item_category_list"].apply(lambda x: x.split(";"))
train_data["item_property_list"] = train_data["item_property_list"].apply(lambda x: x.split(";"))
categories = train_data["predict_category_property"].apply(lambda x: x.split(";"))
train_data["num_query_cat"] = categories.apply(lambda x: len(x))
for i in range(categories.apply(lambda x: len(x)).max()):
train_data["category_"+str(i)] = categories.apply(lambda x: x[i].split(":")[0] if len(x)>i else "-1")
train_data["category_"+str(i)+"_props"] = categories.apply(lambda x: x[i].split(":")[1].split(",") if len(x)>i and x[i].split(":")[0] != "-1" else ["-1"])
#start_time='25'
#end_time='26'
#train_data=train_data[(train_data.day >= int(start_time)) & (train_data.day <int(end_time))]
train_data["num_item_category"] = train_data["item_category_list"].apply(lambda x: len(x)-x.count("-1"))
train_data["num_item_property"] = train_data["item_property_list"].apply(lambda x: len(x)-x.count("-1"))
#####################################
##统计各类别在此次出现前的count数
def count_cat_prep(df,column,newcolumn):
count_dict = {}
df[newcolumn] = 0
data = df[[column,newcolumn]].values
for cat_list in data:
if cat_list[0] not in count_dict:
count_dict[cat_list[0]] = 0
cat_list[1] = 0
else:
count_dict[cat_list[0]] += 1
cat_list[1] = count_dict[cat_list[0]]
df[[column,newcolumn]] = data
train_data['user_item_id'] = train_data['user_id'].astype(str)+"_"+train_data['item_id'].astype(str)
train_data['user_shop_id'] = train_data['user_id'].astype(str)+"_"+train_data['shop_id'].astype(str)
train_data['user_brand_id'] = train_data['user_id'].astype(str)+"_"+train_data['item_brand_id'].astype(str)
train_data['item_category'] = train_data['item_category_list'].astype(str)
train_data['user_category_id'] = train_data['user_id'].astype(str)+"_"+train_data['item_category'].astype(str)
train_data['user_context_id'] = train_data['user_id'].astype(str)+"_"+train_data['context_id'].astype(str)
train_data['user_city_id'] = train_data['user_id'].astype(str)+"_"+train_data['item_city_id'].astype(str)
##统计各类别在总样本中的count数
for column in ['user_id','item_id','item_brand_id','shop_id','user_item_id','user_shop_id','user_brand_id','user_category_id','context_id','item_city_id','user_context_id','user_city_id']:
count_cat_prep(train_data,column,column+'_click_count_prep')
for column in ['user_id','item_id','item_brand_id','shop_id','user_item_id','user_shop_id','user_brand_id','user_category_id','context_id','item_city_id','user_context_id','user_city_id']:
train_data = train_data.join(train_data[column].value_counts(),on = column ,rsuffix = '_count')
print('gen_gaptime ...')
##前一次或后一次点击与现在的时间差（trick）
def lasttime_delta(column):
train_data[column+'_lasttime_delta'] = 0
data = train_data[['context_timestamp',column,column+'_lasttime_delta']].values
lasttime_dict = {}
for df_list in data:
if df_list[1] not in lasttime_dict:
df_list[2] = -1
lasttime_dict[df_list[1]] = df_list[0]
else:
df_list[2] = df_list[0] - lasttime_dict[df_list[1]]
lasttime_dict[df_list[1]] = df_list[0]
train_data[['context_timestamp',column,column+'_lasttime_delta']] = data
def nexttime_delta(column):
train_data[column+'_nexttime_delta'] = 0
data = train_data[['context_timestamp',column,column+'_nexttime_delta']].values
nexttime_dict = {}
for df_list in data:
if df_list[1] not in nexttime_dict:
df_list[2] = -1
nexttime_dict[df_list[1]] = df_list[0]
else:
df_list[2] = nexttime_dict[df_list[1]] - df_list[0]
nexttime_dict[df_list[1]] = df_list[0]
train_data[['context_timestamp',column,column+'_nexttime_delta']]= data
for column in ['user_id','item_id','item_brand_id','shop_id','user_item_id','user_shop_id','user_brand_id','user_category_id','context_id','item_city_id','user_context_id','user_city_id']:
lasttime_delta(column)
train_data = train_data.sort_values('context_timestamp',ascending=False)
for column in ['user_id','item_id','item_brand_id','shop_id','user_item_id','user_shop_id','user_brand_id','user_category_id','context_id','item_city_id','user_context_id','user_city_id']:
nexttime_delta(column)
train_data = train_data.sort_values('context_timestamp')
a_prep=train_data[['instance_id','user_id_click_count_prep','item_id_click_count_prep','item_brand_id_click_count_prep','shop_id_click_count_prep','user_item_id_click_count_prep','user_shop_id_click_count_prep',
'user_brand_id_click_count_prep','user_category_id_click_count_prep','context_id_click_count_prep','item_city_id_click_count_prep','user_context_id_click_count_prep','user_city_id_click_count_prep']]
a_count=train_data[['instance_id','user_id_count','item_id_count','item_brand_id_count','shop_id_count','user_item_id_count','user_shop_id_count','user_brand_id_count','user_category_id_count',
'context_id_count','item_city_id_count','user_context_id_count','user_city_id_count']]
a_instance=train_data['instance_id']
a_gap_time=train_data[['user_id_lasttime_delta','item_id_lasttime_delta','item_brand_id_lasttime_delta','shop_id_lasttime_delta','user_item_id_lasttime_delta',
'user_shop_id_lasttime_delta','user_brand_id_lasttime_delta','user_category_id_lasttime_delta','context_id_lasttime_delta','item_city_id_lasttime_delta',
'user_context_id_lasttime_delta','user_city_id_lasttime_delta',
'user_id_nexttime_delta','item_id_nexttime_delta','item_brand_id_nexttime_delta','shop_id_nexttime_delta','user_item_id_nexttime_delta',
'user_shop_id_nexttime_delta','user_brand_id_nexttime_delta','user_category_id_nexttime_delta','context_id_nexttime_delta','item_city_id_nexttime_delta',
'user_context_id_nexttime_delta','user_city_id_nexttime_delta',
]]