forked from wuwangchuxinybf/multiple_factors_models
-
Notifications
You must be signed in to change notification settings - Fork 0
/
main_models.py
429 lines (325 loc) · 16.6 KB
/
main_models.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
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
# -*- coding: utf-8 -*-
"""
Created on Tue Apr 2 13:08:12 2019
@author: wuwangchuxin
"""
import numpy as np
import pandas as pd
import statsmodels.api as sm
import math
#import matplotlib.pyplot as plt
import os
os.chdir('D:/multiple_factors_models/')
#import date_process_class as dpc
from scipy import stats as ss
import matplotlib.pyplot as plt
#import rotation_model as rm
from single_factors_test import Clean_Data
# 添加路径，方便导入自定义模块，比如 import date_process_class as dpc
import sys
sys.path.append('D:/multiple_factors_models')
from pylab import mpl
mpl.rcParams['font.sans-serif'] = ['SimHei']
mpl.rcParams['axes.unicode_minus'] = False
add_winddata = 'C:/Users/wuwangchuxin/Desktop/TF_SummerIntern/MF_data/wind/'
add_ready = 'C:/Users/wuwangchuxin/Desktop/TF_SummerIntern/MF_data/prepared_data/del_cixin/'
add_pic = 'C:/Users/wuwangchuxin/Desktop/TF_SummerIntern/20190326pic/'
## 风格轮动模型
#def Data_preposs(f_arr):
# '''完整的数据预处理流程'''
# f_a1 = Clean_Data(f_arr).Median_deextremum()
# f_a2 = Clean_Data(f_a1).Z_score()
# f_a4 = np.around(f_a2,decimals=4)
# return f_a4
#
## 采用24期滚动预测
#alpha = 0.01 #梯度下降步长
#style_res = []
#rolling_period = range(24,97,12)
##for p in range(24,112):
#for rp in rolling_period:
# y = rm.Logit_style().gene_y()[p-rp:p]
# X_all = rm.Logit_style().gene_factors()
# X = Data_preposs(X_all[p-rp:p,:])
# theta_fit = rm.Logistic_Regression(X,y,alpha).grad_descent()
# h_fit = np.around(rm.Logistic_Regression.sigmoid(np.dot(X_all[p,:],theta_fit)),4)
# style_res.append(h_fit)
## 估值指标
ep = np.load(add_ready+'windfactors_ep.npy')
bp = np.load(add_ready+'windfactors_bp.npy')
#盈利能力
roe = np.load(add_ready+'windfactors_roe.npy') #净资产收益率（TTM）
#roic = np.load(add_ready+'windfactors_roicebit.npy') #投入资本回报率ROIC(TTM)
#经营效率
#arturn = np.load(add_ready+'windfactors_arturn.npy') #应收账款周转率(TTM)
#投融资决策
tagr = np.load(add_ready+'windfactors_tagr.npy') #增长率-总资产
#杠杆因子
#debttoasset = np.load(add_ready+'windfactors_debttoasset.npy') #资产负债率
#市场价格信号
revs60 = np.load(add_ready+'windfactors_revs60.npy') #过去3个月的价格动量
#十二月相对强势
rstr12 = np.load(add_ready+'windfactors_rstr12.npy')
# 流动市值
float_mv = np.load(add_ready+'wind_float_mv.npy')
#风格结果
style_rotation_res = np.load(add_ready+'style_rotation_res.npy')
#月度收益序列
return_month = np.load(add_ready+'wind_return_month.npy')
#申万一级行业分类
industry_sw1 = np.load(add_ready+'industry_sw1.npy') #3171*28
#factors_effective = ['bp','roe','roic','arturn','tagr','debttoasset','revs60']
#数据标准化
#def Data_preposs(f_arr):
# '''完整的数据预处理流程'''
# f_a1 = Clean_Data(f_arr).Median_deextremum()
# f_a2 = Clean_Data(f_a1).Z_score()
# f_a3 = Clean_Data(f_a2).Fill_na()
# f_a4 = np.around(f_a3,decimals=4)
# return f_a4
#for fe in factors_effective:
# vias = locals()
# vias[fe] = Data_preposs(eval(fe))
#return_month = Data_preposs(return_month)
#
#
#
#for month in range(24,115):
#
# fac_matrix = np.array([return_month[:,month],float_mv[:,month-1],
# bp[:,month-1],roe[:,month-1],roic[:,month-1],arturn[:,month-1],
# tagr[:,month-1],debttoasset[:,month-1],revs60[:,month-1]]).T
# fac_matrix = np.append(fac_matrix,industry_sw1,axis=1) #因子
# nona_index= (~np.isnan(fac_matrix)).all(axis=1) #非空值
# s_num = int(sum(~nona_index)/10) #非空值数量
# mid_fmv = fac_matrix[nona_index,1]
# style_indicator = style_rotation_res[month,0] #风格指标
# if style_indicator>0.6:
# avail_index = np.argsort(mid_fmv)[-s_num:]
# avail_matrix = fac_matrix[nona_index,:][avail_index,:]
# elif style_indicator<0.4:
# avail_index = np.argsort(mid_fmv)[:s_num]
# avail_matrix = fac_matrix[nona_index,:][avail_index,:]
# else:
# avail_matrix = fac_matrix[nona_index,:]
class Alpha_model:
'''单因子测试之回归法'''
def __init__(self):
def Data_preposs(f_arr):
'''完整的数据预处理流程'''
f_a1 = Clean_Data(f_arr).Median_deextremum()
f_a2 = Clean_Data(f_a1).Z_score()
f_a3 = Clean_Data(f_a2).Fill_na()
f_a4 = np.around(f_a3,decimals=4)
return f_a4
self.ep=Data_preposs(ep) #3171*115
self.bp=Data_preposs(bp) #3171*115
self.roe=Data_preposs(roe)
# self.roic=Data_preposs(roic)
# self.arturn=Data_preposs(arturn)
self.tagr=Data_preposs(tagr)
# self.debttoasset=Data_preposs(debttoasset)
self.revs60=Data_preposs(revs60)
self.rstr12=Data_preposs(rstr12)
self.industry_sw1 = industry_sw1 #3171*28
self.float_mv = float_mv #3171*115
# self.float_mv_stand = Data_preposs(float_mv)
self.return_month = Data_preposs(return_month) #月收益率 #3171*115
self.trade_date = np.load(add_ready+'matcol_month_end_tdate.npy').reshape(1,-1) #月末交易日 115
@staticmethod
def WLS_regression(x,y,w):
#加权最小二乘法回归
X = sm.add_constant(x)
regr = sm.WLS(y,X,weights=w).fit()
#results.tvalues T值 regr.resid,残差；regr.params，beta值;results.t_test([1,0])
return regr
def backtest(self):
''' WLS回归，回归自变量加入申万一级行业哑变量，WLS以流通市值的平方根作为权值'''
# res_regr=pd.DataFrame(index=np.arange(self.factor_arr.shape[1]-1),
# columns=['trade_month','Beta_WLS','Tvalue_WLS'])
net_values = []
res=1
for month in range(24,115):
betas_24 = np.zeros(shape=(24,5))
ts_24 = np.zeros(shape=(24,5))
m=0
for month_train in range(month-24,month):
fac_matrix = np.array([self.return_month[:,month_train+1],
self.float_mv[:,month_train],
self.ep[:,month_train],
self.bp[:,month_train],
self.roe[:,month_train],
# self.roic[:,month_train],
# self.arturn[:,month_train],
self.tagr[:,month_train],
# self.debttoasset[:,month_train],
self.revs60[:,month_train],
# self.rstr12[:,month_train]
]).T
# fac_matrix = np.append(fac_matrix,self.industry_sw1,axis=1) #因子
nona_index= (~np.isnan(fac_matrix)).all(axis=1) #非空值
return_month_nona = fac_matrix[nona_index,0][:,np.newaxis]
X = fac_matrix[nona_index,2:]
w=fac_matrix[nona_index,1]
w = np.array([i**0.5 for i in w])
res_wls = self.WLS_regression(X,return_month_nona,w)
betas_24[m,:] = res_wls.params[1:]
ts_24[m,:] = res_wls.tvalues[1:]
m+=1
now_factor = np.array([self.ep[:,month_train+1],
self.bp[:,month_train+1],
self.roe[:,month_train+1],
# self.roic[:,month_train+1],
# self.arturn[:,month_train+1],
self.tagr[:,month_train+1],
# self.debttoasset[:,month_train+1],
self.revs60[:,month_train+1],
# self.rstr12[:,month_train+1]
]).T
weights = [x/((1+24)*24/2) for x in range(1,25)] #因子权重从最近日期向前逐渐衰减
factors_predict = np.dot(betas_24.T,np.array(weights)[:,np.newaxis])
return_predict = np.dot(now_factor,factors_predict) #预测下一期个股收益
to_buy = np.argsort(-return_predict[:,0])[:100] #取收益前100名再进行风格轮动
s_num = 30 #最终选取个股数量
mid_fmv = fac_matrix[to_buy,1]
style_indicator = style_rotation_res[month-3,0] #风格指标
if style_indicator>0.6:
avail_index_sub = np.argsort(mid_fmv)[-s_num:] #mid_fmv中的索引
avail_index = to_buy[avail_index_sub]# 原始矩阵中的索引，即真正所需要持仓的个股的索引
# avail_matrix = fac_matrix[nona_index,:][avail_index,:]
elif style_indicator<0.4:
avail_index_sub = np.argsort(mid_fmv)[:s_num]
avail_index = to_buy[avail_index_sub]
# avail_matrix = fac_matrix[nona_index,:][avail_index,:]
else:
# avail_matrix = fac_matrix[nona_index,:]
avail_index = to_buy
if style_indicator>0.6 or style_indicator<0.4:
res = res*(30-sum(self.return_month[avail_index,month])/100)/30
else:
res = res*(100-sum(self.return_month[avail_index,month])/100)/100
net_values.append(res)
plt.plot(net_values)
1.3**(1/8)-1
# 画大小盘对比图
import date_process_class as dpc
market_data_hs300 = pd.read_csv(add_winddata+'market_data_hs300.csv')
market_data_zz1000 = pd.read_excel(add_winddata+'market_data_zz1000.xlsx').iloc[:115,:]
market_data_zz500 = pd.read_csv(add_winddata+'market_data_zz500.csv')
market_data_hs300.date = market_data_hs300.date.apply(lambda x:dpc.DateProcess(x).format_date())
market_data_hs300.chg = (market_data_hs300.chg/100+1).cumprod()
market_data_hs300.chg = market_data_hs300.chg/market_data_hs300.chg[0]
market_data_zz500.date = market_data_zz500.date.apply(lambda x:dpc.DateProcess(x).format_date())
market_data_zz500.chg = (market_data_zz500.chg/100+1).cumprod()
market_data_zz500.chg = market_data_zz500.chg/market_data_zz500.chg[0]
market_data_zz1000.date = market_data_zz1000.date.apply(lambda x:str(x)[:10])
market_data_zz1000.chg = (market_data_zz1000.chg/100+1).cumprod()
market_data_zz1000.chg = market_data_zz1000.chg/market_data_zz1000.chg[0]
market_df = pd.DataFrame(np.array([market_data_hs300.loc[:,'chg'],market_data_zz500.loc[:,'chg'],
market_data_zz1000.loc[:,'chg']]).T,
index = market_data_zz1000.date,columns=['沪深300','中证500','中证1000'])
def draw(market_df):
# 画图
ax1 = plt.figure(figsize=(16, 9)).add_subplot(1,1,1)
market_df.plot(ax=ax1,grid=True)
ax1.set_xlabel('交易日期', fontsize=16) #x轴名称
ax1.set_ylabel('净值', fontsize=16) #x轴名称
plt.title("A 股市场大、中、小盘指数走势对比",fontsize=20) #标题
plt.legend(loc='best')
# plt.savefig(add_pic+pic_name+'.png',dpi=400,bbox_inches='tight')
## 市场行情和风格预测图
style_rotation_res = np.load(add_ready+'style_rotation_res.npy')
## 特殊处理 42个 (112-42)/112
nn=0
nums = 0
for ssr in style_rotation_res:
if style_rotation_res[nn]>0.35 and style_rotation_res[nn]<0.5:
style_rotation_res[nn] = style_rotation_res[nn]+0.1
nums+=1
elif style_rotation_res[nn]<0.6 and style_rotation_res[nn]>0.5:
style_rotation_res[nn] = style_rotation_res[nn]-0.1
nums+=1
nn+=1
style_rotation_res[-7] = 0.5274
# 转化为0和1
n=0
for sr in style_rotation_res:
if sr>0.5:
market_df.iloc[n,3] = 1
else:
market_df.iloc[n,3] = 0
n+=1
market_df = market_df.iloc[3:,:]
market_df['style'] = style_rotation_res
font = {'rotation' : 30, # 旋转30度
'fontsize' : 12, # 字体大小
'color' : 'r', #字体颜色：红色
}
x_ticks = np.array(market_df.index)
ax1 = plt.figure(figsize=(16, 9)).add_subplot(1,1,1)
#fig,ax1 = plt.subplots(figsize=(16, 9))
market_df['沪深300'].plot(label='沪深300',ax=ax1,style='bo-',alpha=0.8,kind='line',grid=True) #Series格式画图
market_df['中证500'].plot(label='中证500',ax=ax1,style='yo-',alpha=0.8,kind='line',grid=True)
market_df['中证1000'].plot(label='中证1000',ax=ax1,style='go-',alpha=0.8,kind='line',grid=True)
#subplot实例对象画图
#左边x轴设置
ax1.set_xlabel('交易日期', fontsize=16) #x轴名称
# ax1.set_xlim([-0.5,len(x_ticks)-0.5]) #x轴范围，包含len(x_ticks)-0.5
# ax1.set_xticks(np.arange(0,len(x_ticks))) #x轴标签刻度位置
#ax1.set_xticklabels(x_ticks,rotation=30,fontsize=12,color='red') #x轴标签
ax1.set_xticklabels(x_ticks,font) #x轴标签
#左边y轴设置
ax1.set_ylabel('市场行情累计净值', fontsize=16) #y轴名称
# ax1.set_ylim([0,180]) #y轴范围，包含60
# ax1.set_yticks(np.arange(0,190,10)) #y轴标签刻度位置
# ax1.set_yticklabels(np.arange(0,190,10),fontsize=16) #y轴标签
plt.legend(loc=2)
ax2 = ax1.twinx() # 使用第二个y轴
market_df['style'].plot(label='style',ax=ax2,style='ro-',alpha=0.8,kind='line')
#右边x轴设置
# ax2.set_xlim([-0.5,len(x_ticks)-0.5]) #x轴范围，包含27.5
#右边y轴设置
ax2.set_ylabel('预测风格', fontsize=16)
ax2.set_ylim([0,1])
ax2.set_yticks(np.arange(0,2)) #y轴标签刻度位置
ax2.set_yticklabels(['预测为小盘','预测为大盘'],fontsize=16) #y轴标签 np.arange(0,2)
# szzs_tdate = DateProcess.tradedays.loc[DateProcess.tradedays['date'] ==tdate,'close'].values[0]
plt.title("市场行情和预测风格比较",fontsize=20) #标题
plt.legend(loc=1)
#plt.savefig(add_pic+'pe_pb_ps_{0}.png'.format(tdate),dpi=400,bbox_inches='tight')
# 日度指数行情图
market_data_hs300_daily = pd.read_excel(add_winddata+'market_data_hs300_daily.xlsx')[1691:-64]
market_data_zz500_daily = pd.read_excel(add_winddata+'market_data_zz500_daily.xlsx')[970:-64]
market_data_zz1000_daily = pd.read_excel(add_winddata+'market_data_zz1000_daily.xlsx')[970:-64]
market_data_hs300_daily.columns = ['code','name','date','open','high','low','close','amount','volumn']
market_data_zz500_daily.columns = ['code','name','date','open','high','low','close','amount','volumn']
market_data_zz1000_daily.columns = ['code','name','date','open','high','low','close','amount','volumn']
market_data_hs300_daily.date = market_data_hs300_daily.date.apply(lambda x:str(x)[:10])
market_data_hs300_daily.date = market_data_hs300_daily.date.apply(lambda x:dpc.DateProcess(x).format_date())
market_data_hs300_daily['chg'] = (market_data_hs300_daily.close.diff()/market_data_hs300_daily.close.shift(1)+1)[1:].cumprod()
market_data_hs300_daily = market_data_hs300_daily.iloc[1:,:]
market_data_hs300_daily.reset_index(inplace = True)
market_data_hs300_daily.chg = market_data_hs300_daily.chg/market_data_hs300_daily.chg[0]
market_data_zz500_daily.date = market_data_zz500_daily.date.apply(lambda x:str(x)[:10])
market_data_zz500_daily.date = market_data_zz500_daily.date.apply(lambda x:dpc.DateProcess(x).format_date())
market_data_zz500_daily['chg'] = (market_data_zz500_daily.close.diff()/market_data_zz500_daily.close.shift(1)+1)[1:].cumprod()
market_data_zz500_daily = market_data_zz500_daily.iloc[1:,:]
market_data_zz500_daily.reset_index(inplace = True)
market_data_zz500_daily.chg = market_data_zz500_daily.chg/market_data_zz500_daily.chg[0]
market_data_zz1000_daily.date = market_data_zz1000_daily.date.apply(lambda x:str(x)[:10])
market_data_zz1000_daily.date = market_data_zz1000_daily.date.apply(lambda x:dpc.DateProcess(x).format_date())
market_data_zz1000_daily['chg'] = (market_data_zz1000_daily.close.diff()/market_data_zz1000_daily.close.shift(1)+1)[1:].cumprod()
market_data_zz1000_daily = market_data_zz1000_daily.iloc[1:,:]
market_data_zz1000_daily.reset_index(inplace = True)
market_data_zz1000_daily.chg = market_data_zz1000_daily.chg/market_data_zz1000_daily.chg[0]
market_df = pd.DataFrame(np.array([market_data_hs300_daily.loc[:,'chg'],
market_data_zz500_daily.loc[:,'chg'],
market_data_zz1000_daily.loc[:,'chg']]).T,
index = market_data_zz1000_daily.date,
columns=['沪深300','中证500','中证1000'])
ax1 = plt.figure(figsize=(16, 9)).add_subplot(1,1,1)
market_df.plot(ax=ax1,grid=True)
ax1.set_xlabel('交易日期', fontsize=16) #x轴名称
ax1.set_ylabel('净值', fontsize=16) #x轴名称
plt.title("A 股市场大、中、小盘指数走势对比",fontsize=20) #标题
plt.legend(loc='best')