-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathbase_feature.py
96 lines (76 loc) · 4.49 KB
/
base_feature.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
import pandas as pd
import numpy as np
from pandas import DataFrame as DF
def get_division_feature(data ,feature_name):
# 创造出特征之间 进行四则变换的特征， 每两两特征之间进行变化， 并且进行变换之后把变化后的 新特证名记录下来，（知道了吧。不想之前直接一起，
# 特征名都完全丢失了
new_feature = []
new_feature_name = []
for i in range(len(data[feature_name].columns ) -1):
for j in range( i +1 ,len(data[feature_name].columns)):
# 保存新创建的特征值和特征名
new_feature_name.append(data[feature_name].columns[i] + '/' + data[feature_name].columns[j])
new_feature_name.append(data[feature_name].columns[i] + '*' + data[feature_name].columns[j])
new_feature_name.append(data[feature_name].columns[i] + '+' + data[feature_name].columns[j])
new_feature_name.append(data[feature_name].columns[i] + '-' + data[feature_name].columns[j])
new_feature.append(data[data[feature_name].columns[i] ] /data[data[feature_name].columns[j]])
new_feature.append(data[data[feature_name].columns[i] ] *data[data[feature_name].columns[j]])
new_feature.append(data[data[feature_name].columns[i] ] +data[data[feature_name].columns[j]])
new_feature.append(data[data[feature_name].columns[i] ] -data[data[feature_name].columns[j]])
temp_data = DF(pd.concat(new_feature ,axis=1))
temp_data.columns = new_feature_name
data = pd.concat([data ,temp_data] ,axis=1).reset_index(drop=True)
print(data.shape)
return data.reset_index(drop=True)
def get_square_feature(data, feature_name):
# 对特征进行二项变换，开放等方式， 并将处理后的特征名记录
new_feature = []
new_feature_name = []
for i in range(len(data[feature_name].columns)):
new_feature_name.append(data[feature_name].columns[i] + '**2')
new_feature_name.append(data[feature_name].columns[i] + '**1/2')
new_feature.append(data[data[feature_name].columns[i]] ** 2)
new_feature.append(data[data[feature_name].columns[i]] ** (1 / 2))
temp_data = DF(pd.concat(new_feature, axis=1))
temp_data.columns = new_feature_name
data = pd.concat([data, temp_data], axis=1).reset_index(drop=True)
print(data.shape)
return data.reset_index(drop=True)
if __name__=='__main__':
train = pd.read_csv( './data/do_data_pred/train.csv')
test = pd.read_csv( './data/do_data_pred/test.csv')
y_target = train.pop('发电量')
################### 1. 转换率、电压、电流、功率再造特征 ###################
#平均数值
# train['平均转换效率']=(train['转换效率A']+train['转换效率B']+train['转换效率C'])/3
# train['平均电压']=(train['电压A']+train['电压B']+train['电压C'])/3
# train['平均电流']=(train['电流A']+train['电流B']+train['电流C'])/3
# train['平均功率']=(train['功率A']+train['功率B']+train['功率C'])/3
#
#
# test['平均转换效率']=(test['转换效率A']+test['转换效率B']+test['转换效率C'])/3
# test['平均电压']=(test['电压A']+test['电压B']+test['电压C'])/3
# test['平均电流']=(test['电流A']+test['电流B']+test['电流C'])/3
# test['平均功率']=(test['功率A']+test['功率B']+test['功率C'])/3
################### 2. 与平均值的差距特征 ###################
# for i in train.columns:
# train['与均值差'+i]=train[i]-np.mean(train[i])
# for i in test.columns:
# test['与均值差'+i]=test[i]-np.mean(test[i])
################### 3. 周期特征（距离波底的距离） ###################
#自造周期特征 计算规则，依据于 光照强度
################### 4. 变幻和交叉特征 ###################
feature_name = [i for i in train.columns]
#feature_SNP = [i for i in feature_name]
#feature_name = [i for i in train_data.columns]
train_data = get_division_feature(train, feature_name)
test_data = get_division_feature(test, feature_name)
temp_data = DF(y_target)
temp_data.columns = ['发电量']
print(train_data.columns)
print('特征间变换')
train_data = get_square_feature(train_data, feature_name)
test_data = get_square_feature(test_data, feature_name)
train = pd.concat([train_data, temp_data], axis=1).reset_index(drop=True)
train.to_csv('./data/do_feat_data_3/train.csv')
test_data.to_csv('./data/do_feat_data_3/test.csv')