-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy path2_lightGBM.py
103 lines (89 loc) · 3.81 KB
/
2_lightGBM.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
# -*- coding: utf-8 -*-
"""
Created on Fri Mar 2 20:40:40 2018
@author: Liaowei
"""
import pandas as pd
import numpy as np
import time
import datetime
import matplotlib.pyplot as plt
from sklearn.cross_validation import KFold,train_test_split
from sklearn.metrics import f1_score
import lightgbm as lgb
import itertools
import pickle
from pylab import mpl
mpl.rcParams['font.sans-serif'] = ['FangSong'] # 指定默认字体
mpl.rcParams['axes.unicode_minus'] = False # 解决保存图像是负号'-'显示为方块的问题
np.seterr(invalid='ignore')
#train_df = pd.read_csv('../data/linear_train.csv',encoding='gb2312')
#unk_test_df = pd.read_csv('../data/linear_test.csv',encoding='gb2312')
train_df = pd.read_csv('../data/unlinear_train.csv',encoding='gb2312')
unk_test_df = pd.read_csv('../data/unlinear_test.csv',encoding='gb2312')
unk_test_Y = pd.read_csv('../data/f_answer_a_20180306.csv',header=None)
train_Y = pickle.load(open('../data/train_Y', 'rb'))
kf = KFold(len(train_df), n_folds = 5, shuffle=True, random_state=520)
#误差函数
def evalerror(pred, df):
label = df.get_label().values.copy()
pred = [1 if i>=0.5 else 0 for i in pred]
score = f1_score(label,pred)
#返回list类型，包含名称，结果，is_higher_better
return ('F1',score,False)
#设置lightGBM的参数
#metric参数没有F1改使用哪个评价指标？
print('开始训练...')
params = {
# 'num_leaves': 30, #70
'max_depth': 3, #7
# 'min_data_in_leaf': 100, #30
'feature_fraction': 0.85 #1
# 'learning_rate': 0.48, #0.48 0.36
#
# 'boosting_type': 'gbdt',
# 'objective': 'binary',
# 'verbose': -1,
# 'metric': 'binary_logloss',
# 'bagging_seed': 3,
}
cv_preds = np.zeros(train_df.shape[0])
unk_test_preds = np.zeros((unk_test_df.shape[0], 5))
#test_preds = np.zeros((test_df.shape[0], 5))
for i, (train_index, cv_index) in enumerate(kf):
print('第{}次训练...'.format(i))
train_feat = train_df.iloc[train_index]
cv_feat = train_df.iloc[cv_index]
lgb_train = lgb.Dataset(train_feat.values, train_Y.loc[train_index])
lgb_cv = lgb.Dataset(cv_feat.values, train_Y.loc[cv_index])
gbm = lgb.train(params=params, #参数
train_set=lgb_train, #要训练的数据
num_boost_round=6000, #迭代次数
valid_sets=lgb_cv, #训练时需要评估的列表
verbose_eval=False, #
feval=evalerror, #误差函数
early_stopping_rounds=500)
#评价特征的重要性
feat_imp = pd.Series(gbm.feature_importance(), index=train_df.columns).sort_values(ascending=False)
cv_preds[cv_index] += gbm.predict(cv_feat.values)
unk_test_preds[:,i] = gbm.predict(unk_test_df.values)
# test_preds[:,i] = gbm.predict(test_df.values)
#看看训练结果
temp_train_preds = gbm.predict(train_df.values)
temp_train_preds = [1 if i>=0.5 else 0 for i in temp_train_preds]
cv_preds = [1 if i>=0.5 else 0 for i in cv_preds]
#test_preds = np.mean(test_preds, axis=1)
#test_preds = [1 if i>=0.5 else 0 for i in test_preds]
print('训练得分:', f1_score(temp_train_preds, train_Y.values))
print('线下得分:', f1_score(cv_preds, train_Y.values))
#print('测试得分:', f1_score(test_preds, test_Y.values))
'''
目前参数训练得分0.855，线下得分0.665，存在很大的过拟合
'''
#测试数据
unk_test_preds = np.mean(unk_test_preds, axis=1)
unk_test_preds = [1 if i>=0.5 else 0 for i in unk_test_preds]
print('线下得分:', f1_score(unk_test_preds, unk_test_Y.values))
#submission = pd.DataFrame({'pred':unk_test_preds})
#submission.to_csv(r'../data/lgb_feat_select{}.csv'.format(datetime.datetime.now().strftime('%Y%m%d_%H%M%S')),
# header=None,index=False, float_format='%.4f')