1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
#导入需要用到的库
import numpy as np
import pandas as pd
from jqfactor import get_factor_values
from jqlib.technical_analysis import *
from jqfactor import neutralize
from jqfactor import winsorize_med
from jqfactor import standardlize
from jqfactor import winsorize
from IPython.core.interactiveshell import InteractiveShell
InteractiveShell.ast_node_interactivity = "all"
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import _validation
from sklearn.metrics import precision_score, recall_score, accuracy_score
from sklearn.metrics import classification_report
from sklearn.metrics import roc_curve, auc
import matplotlib.pyplot as plt
import time
from sklearn.model_selection import GridSearchCV
from sklearn.ensemble import RandomForestClassifier
from sklearn.cross_validation import train_test_split

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
print 'Starting yudong-back-test, please wait ...........' 
start_time = time.time()
stocklist = get_index_stocks('000300.XSHG')   

sample_1 = pd.read_csv('data/data_train_2007to2010.csv')
sample_2 = pd.read_csv('data/data_train_2007to2011.csv')
sample_3 = pd.read_csv('data/data_train_2007to2012.csv')
sample_4 = pd.read_csv('data/data_train_2008to2013.csv')
sample_5 = pd.read_csv('data/data_train_2009to2014.csv')
sample_6 = pd.read_csv('data/data_train_2010to2015.csv')
sample_7 = pd.read_csv('data/data_train_2011to2016.csv')
sample_8 = pd.read_csv('data/data_train_2012to2017.csv')
sample2011to2017 = [sample_1, sample_2, sample_3, sample_4, sample_5, sample_6, sample_7]


year_start = pd.to_datetime(pd.date_range('2011-01-01','2017-12-31',freq = 'AS'))
year_end = pd.to_datetime(pd.date_range('2011-01-01','2017-12-31',freq = 'A'))

accuracy_test_list = [] 
precision_test_list = []
auc_test_list = []
recall_test_list = []
factor_score_list = []
gs_list = []
year = 2011
    
for i,j,k in zip(year_start, year_end,sample2011to2017):
    print '%s年回测开始：............'%year
    gs,a,b,c,d,factor_score = test_out(k, i,j,stocklist)
    gs_list.append(gs)
    factor_score_list.append(factor_score)
    accuracy_test_list += a
    precision_test_list += b
    auc_test_list += c
    recall_test_list += d
    print '%s年回测已经结束了............'%year
    print '\n'
    year += 1
    

1
2
3
4
5
6
7
8
gs_last, e,f,g,h,factor_score_last = test_out(sample_8, '2018-01-01', '2018-08-15',stocklist)
accuracy_test_list += e
precision_test_list += f
auc_test_list += g
recall_test_list += h
gs_list.append(gs_last)
factor_score_list.append(factor_score_last)
print 'Total cost time: %s'%(time.time() - start_time)

1
2
3
4
5
#打印每年因子重要性
factor_importance = factor_score_list[0]
for i in factor_score_list[1:]:
    factor_importance = pd.merge(factor_importance, i, left_on = 'features',right_on = 'features')
factor_importance

1
2
factor_importance.columns = ['features']+range(2011,2019)
factor_importance

1
2
3
4
average(auc_test_list)
average(precision_test_list)
average(accuracy_test_list)
average(recall_test_list)

1
2
3
4
draw(accuracy_test_list,'Accuracy','b')
draw(precision_test_list,'Precision','g')
draw(auc_test_list,'AUC','r')
draw(recall_test_list,'Recall','c')