forked from out-to-right/--rong360
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Clear_user_info.py
194 lines (156 loc) · 7.48 KB
/
Clear_user_info.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
# coding=utf-8
'''
author: ShiLei Miao
'''
from numpy import *
import pandas as pd
from pandas import *
import os
import time
os.chdir(r'E:\PycharmProjects\Rong360')
start =time.clock()
###
# 数据读取
user_info = pd.read_csv('***\\user_info.txt')
u_train = pd.read_csv('***\\train.txt')
u_test = pd.read_csv('***\\test.txt')
##### Step1:
### user_info表处理
# 大量数据重复，目前考虑将变量‘tm_encode’踢出来，将user_info表拆分
# user_info1 96153记录
user_info1 = user_info.drop(['tm_encode'],axis=1)
user_info1 = user_info1.drop_duplicates()
# user_info2 38261记录【train:26000, test:12261】
user_info['tm_encode'] = user_info['tm_encode'].astype(float)
N_user_info2 = DataFrame()
N_user_info2['min_tm_encode'] = user_info.groupby('user_id')['tm_encode'].min()
N_user_info2['max_tm_encode'] = user_info.groupby('user_id')['tm_encode'].max()
N_user_info2['sum_tm_encode'] = user_info.groupby('user_id')['tm_encode'].sum()
N_user_info2['mean_tm_encode'] = user_info.groupby('user_id')['tm_encode'].mean()
N_user_info2['std_tm_encode'] = user_info.groupby('user_id')['tm_encode'].std()
N_user_info2['count_tm_encode'] = user_info.groupby('user_id')['tm_encode'].count()
N_user_info2['count_user_id'] = user_info.groupby('user_id')['user_id'].count()
##### Step2:
#...................................................user_info2已处理完毕
### user_info1表处理
# 表中记录按照是否为38261行将user_info1表拆分【结果：Uni_dta为10】
Uni_dta = []
U_Uni_dta = []
U_Uni_dta.append('user_id')
L_dta = DataFrame()
for i in user_info1.columns:
L_dta['user_id'] = user_info1['user_id']
L_dta[i] = user_info1[i]
L_dta = L_dta.drop_duplicates()
if L_dta.shape[0] == 38261:
Uni_dta.append(i)
if L_dta.shape[0] != 38261:
U_Uni_dta.append(i)
#print i
#print L_dta.shape[0]
#print '\n'
L_dta = DataFrame()
# 记录为38261行
user_info1_1 = DataFrame(user_info1,columns=Uni_dta)
N_user_info1_1 = user_info1_1.drop_duplicates()
N_user_info1_1 = DataFrame(N_user_info1_1.values,columns=N_user_info1_1.columns)
# N_user_info1_1的缺失处理方式【mode】
def MissData_deal_N_user_info1_1(data):
for i in range(data.shape[1]):
if data.ix[:, i].isnull().sum() >= 1:
# 保留数据的缺失信息
data[data.columns[i] + '_missingInfo'] = data.ix[:,i].isnull().astype(int)
# 类别变量用众数填充
data[data.columns[i]] = data[data.columns[i]].fillna(data.ix[:, i].mode()[0])
return data
N_user_info1_1 = MissData_deal_N_user_info1_1(N_user_info1_1)
# 记录非38261行
user_info1_2 = DataFrame(user_info1,columns=U_Uni_dta)
user_info1_2 = user_info1_2.drop_duplicates()
##### Step3:
#...................................................user_info1_1已处理完毕
### user_info1_2表处理
# 按表中变量的性质拆分为：user_info1表拆分user_info1_2_1【类别型】与user_info1_2_2【数值型】
N_var1 = ['user_id','local_hk','money_function','sex','occupation','education','marital_status','live_info']
N_var2 = ['user_id','age','expect_quota','max_month_repay','salary']
def MissData_deal_user_info1_2_1(data):
for i in range(data.shape[1]):
if data.ix[:, i].isnull().sum() >= 1:
# 保留数据的缺失信息
data[data.columns[i] + '_missingInfo'] = data.ix[:,i].isnull().astype(int)
# 类别变量用-1填充
data[data.columns[i]] = data[data.columns[i]].fillna(-1)
return data
user_info1_2_1 = DataFrame(user_info1_2,columns=N_var1)
user_info1_2_1 = MissData_deal_user_info1_2_1(user_info1_2_1)
def Dummy(data, variable):
One_hot_dta = get_dummies(data, prefix=variable)
return One_hot_dta
def Dummy_Master(data):
names = data.columns
for i in names:
if i != "user_id" and len(data[i].value_counts()) > 2:
One_hot_dtas = Dummy(data[i], i)
data = data.join(One_hot_dtas)
data = data.drop(i, axis=1)
return data
# 将类别变量进行one_hot编码
user_info1_2_1 = Dummy_Master(user_info1_2_1)
# 将编码后的变量按照频数统计
N_user_info1_2_1 = DataFrame()
for i in range(1,user_info1_2_1.shape[1]):
Var_N = user_info1_2_1.columns[i]
N_user_info1_2_1[Var_N] = user_info1_2_1.groupby('user_id')[Var_N].count()
##### Step4:
#...................................................user_info1_2_1已处理完毕
### user_info1_2_2表处理
# 按表中变量统计相关的指标
user_info1_2_2 = DataFrame(user_info1_2,columns=N_var2)
N_user_info1_2_2_1 = DataFrame()
N_111 = DataFrame(user_info1_2_2[user_info1_2_2['age']!='NONE'],\
columns=['user_id','age']).drop_duplicates()
N_111['age'] = N_111['age'].astype(float)
N_user_info1_2_2_1['age'] = N_111.groupby('user_id')['age'].mean()
user_info1_2_2_2 = user_info1_2_2.drop(['age'],axis=1)
def MissData_deal_Num(data):
for i in range(data.shape[1]):
if data.ix[:, i].isnull().sum() >= 1:
# 保留数据的缺失信息
data[data.columns[i] + '_missingInfo'] = data.ix[:,i].isnull().astype(int)
# 类别变量用均值填充
#data[data.columns[i]] = data[data.columns[i]].fillna(data.ix[:, i].mean())
return data
user_info1_2_2_2 = MissData_deal_Num(user_info1_2_2_2)
N_user_info1_2_2_2 = DataFrame()
for i in range(1,user_info1_2_2_2.shape[1]):
if i < 4:
Var_N = user_info1_2_2_2.columns[i]
N_user_info1_2_2_2[Var_N+'_max'] = user_info1_2_2_2.groupby('user_id')[Var_N].max()
N_user_info1_2_2_2[Var_N+'_min'] = user_info1_2_2_2.groupby('user_id')[Var_N].min()
N_user_info1_2_2_2[Var_N+'_mean'] = user_info1_2_2_2.groupby('user_id')[Var_N].mean()
N_user_info1_2_2_2[Var_N+'_std'] = user_info1_2_2_2.groupby('user_id')[Var_N].std()
N_user_info1_2_2_2[Var_N+'_count'] = user_info1_2_2_2.groupby('user_id')[Var_N].count()
if i > 3:
Var_N = user_info1_2_2_2.columns[i]
N_user_info1_2_2_2[Var_N] = user_info1_2_2_2.groupby('user_id')[Var_N].count()
N_user_info1_2_2_2 = N_user_info1_2_2_2.fillna(-1)
##### 数据集的合并
##..........................训练集...........
N_user_info2['user_id'] = N_user_info2.index
N_user_info1_2_1['user_id'] = N_user_info1_2_1.index
N_user_info1_2_2_1['user_id'] = N_user_info1_2_2_1.index
N_user_info1_2_2_2['user_id'] = N_user_info1_2_2_2.index
Master_one = merge(u_train, N_user_info2, how="left", left_on='user_id', right_on='user_id')
Master_two = merge(Master_one, N_user_info1_1, how="left", left_on='user_id', right_on='user_id')
Master_three = merge(Master_two, N_user_info1_2_1, how="left", left_on='user_id', right_on='user_id')
Master_four = merge(Master_three, N_user_info1_2_2_1, how="left", left_on='user_id', right_on='user_id')
N_train_user_info = merge(Master_four, N_user_info1_2_2_2, how="left", left_on='user_id', right_on='user_id')
Master_one = merge(u_test, N_user_info2, how="left", left_on='user_id', right_on='user_id')
Master_two = merge(Master_one, N_user_info1_1, how="left", left_on='user_id', right_on='user_id')
Master_three = merge(Master_two, N_user_info1_2_1, how="left", left_on='user_id', right_on='user_id')
Master_four = merge(Master_three, N_user_info1_2_2_1, how="left", left_on='user_id', right_on='user_id')
N_test_user_info = merge(Master_four, N_user_info1_2_2_2, how="left", left_on='user_id', right_on='user_id')
N_train_user_info.to_csv('***\\N_train_user_info.csv',index=False)
N_test_user_info.to_csv('***\\N_test_user_info.csv',index=False)
end = time.clock()
print ('Running time: %s Seconds'%(end-start))