-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathPipeline_single.py
126 lines (93 loc) · 4.65 KB
/
Pipeline_single.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
# coding: utf-8
from QSAR_package.data_split import extractData,randomSpliter
from QSAR_package.feature_preprocess import correlationSelection,RFE_ranking
from QSAR_package.data_scale import dataScale
from QSAR_package.grid_search import gridSearchPlus,gridSearchBase
from QSAR_package.model_evaluation import modeling
import pandas as pd
import numpy as np
import os
os.system("")
if __name__ == '__main__':
"""=============================参数设定（常变）================================"""
# ##输入文件路径，包含所有样本的描述符数据文件
file_name = "./spla2_296_rdkit2d.csv"
# ##标签列的列名，label列请务必放于第一个feature列的前一列
label_name = 'pIC50'
# ##随机划分训练集和测试集的随机种子
randx = 0
# ##描述符权重排序方法，可选["Pearson","SVM-RFE","RF-RFE"]
ranker = 'Pearson'
# ##建模算法，可选["SVC","DTC","RFC","SVR","RFR"]，也可直接传入一个学习器对象
estimatorName = 'SVR'
# ##网格寻优中交叉验证的重数
gridCV = 3
# ##网格寻优的重复次数
repeat = 3
# ##触发early_stop的最小mean_test_score变化率，如果设为0，则不执行early_stop
early_stop = 0.01
# ##采用带描述符数量的重复网格寻优，指定描述符数量的迭代范围，如果不迭代描述符数量，则删除或注释掉此参数
# features_range = (19,20)
# ##是否做散点图，回归任务可以设为True，分类任务请设为False
make_fig = False
"""========================参数设定（不常变）=========================="""
# ##测试集的划分比例
test_size = 0.25
# ##Pearson相关性筛选中描述符与活性相关性的下界，默认值为0.1
corr_low = 0.1
# ##Pearson相关性筛选中描述符之间两两相关性的上界，默认值为0.9
corr_up = 0.9
# ##连续型特征数据的压缩范围
scale_range = (0.1,0.9)
# ##最后拟合的模型做交叉验证的重数，可以是整数，或字符串"LOO"，不做cv可设为0
make_cv = 5
# ##是否保存最终模型文件
save_model = True
"""========输入总描述符CSV文件→随机分集→相关性筛选(排序)→数据压缩 [ →RFE排序 ]=========="""
spliter = randomSpliter(test_size=test_size,random_state=randx)
spliter.ExtractTotalData(file_name,label_name=label_name)
spliter.SplitData()
tr_x = spliter.tr_x
tr_y = spliter.tr_y
te_y = spliter.te_y
corr = correlationSelection()
corr.PearsonXX(tr_x, tr_y,threshold_low=corr_low, threshold_up=corr_up)
scaler = dataScale(scale_range=scale_range)
tr_scaled_x = scaler.FitTransform(corr.selected_tr_x.iloc[:,:])
te_scaled_x = scaler.Transform(spliter.te_x,DataSet='test')
if ranker == 'Pearson':
tr_ranked_x = tr_scaled_x
te_ranked_x = te_scaled_x
elif ranker == 'SVM-RFE':
if np.unique(tr_y) <= 5:
rfe = RFE_ranking('SVC',features_num=1)
else:
rfe = RFE_ranking('SVR',features_num=1)
rfe.Fit(tr_scaled_x, tr_y)
tr_ranked_x = rfe.tr_ranked_x
te_ranked_x = te_scaled_x.loc[:,tr_ranked_x.columns]
elif ranker == 'RF-RFE':
if np.unique(tr_y) <= 5:
rfe = RFE_ranking('RFC',features_num=1)
else:
rfe = RFE_ranking('RFR',features_num=1)
rfe.Fit(tr_scaled_x, tr_y)
tr_ranked_x = rfe.tr_ranked_x
te_ranked_x = te_scaled_x.loc[:,tr_ranked_x.columns]
"""=========repeat寻优，找出最佳参数 [ 及描述符数量 ]============="""
grid = gridSearchPlus(grid_estimatorName=estimatorName, fold=gridCV, repeat=repeat, early_stop=early_stop, scoreThreshold=None)
if('features_range' in vars()):
grid.FitWithFeaturesNum(tr_ranked_x, tr_y,features_range=features_range) # 用带描述符数量的寻优策略
else:
grid.Fit(tr_scaled_x,tr_y) # 不带描述符数量的寻优
"""==============拟合模型，评价模型，保存结果================"""
model = modeling(grid.best_estimator,params=grid.best_params)
model.Fit(tr_scaled_x.loc[:,grid.best_features], tr_y)
model.Predict(te_scaled_x.loc[:,grid.best_features],te_y)
if make_cv:
model.CrossVal(cv=make_cv)
model.ShowResults(show_cv=True, make_fig=make_fig)
else:
model.ShowResults(show_cv=False, make_fig=make_fig)
model.SaveResults(file_name[:-4]+'_results.csv',notes='{},split_seed={},gridCV={}'.format(ranker,randx,gridCV),
save_model=save_model)