-
Notifications
You must be signed in to change notification settings - Fork 0
/
lstm_cpam.py
247 lines (216 loc) · 10.4 KB
/
lstm_cpam.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
'''
主算法,跑lstm和cpam
'''
import tushare as ts
import numpy as np
import pandas as pd
import datetime
from datetime import timedelta, date
from predict import CPAM
import warnings
warnings.filterwarnings('ignore')
sh_a_df = pd.read_csv("sh_A.csv") # 上证指数
# 设置token
token= '6acaa6d1a6872945fba43e7418f7b9f47d2eee0ab1aeaf63b5ec3ec3'
ts.set_token(token)
pro=ts.pro_api()
# 读取df
df = pd.read_csv("Process_Data_5_tuple.csv",encoding = 'gbk')
df_company_class = pd.read_excel("上证公司数据.xlsx")
# 设置窗口获取时间--> 到访前窗口[before_window,date], 到访后窗口[date,after_window],收集的数据窗口[time_from,before_window]
before_window = 35
after_window = 40
time_from = 182
before_limit = 20+1
after_limit = 25
print("公司数量为:",len(df))
df_array = np.array(df)
# print(df_array)
# 得到AR的累计值
def get_ar_sum(list):
ar_sum_list = []
tmp = 0
for i in range(len(list)):
tmp+=list[i]
ar_sum_list.append(tmp)
return ar_sum_list
def match_index(df,choice):
# choice 表示不排序,为-1代表逆序,为1代表正序
df["index_pct_chg"] = 0.0000
df["number"] = 0
# if(choice==1):
# print("长度是:",df.shape[0])
for i in range(0,df.shape[0]):
# if(choice==1):
# print(i)
if(choice==-1):
# print("111")
df["number"][i] = -i
if(choice==1):
df["number"][i] = i+1
# print(df["number"][i])
temp = df['trade_date'][i]
if (1 <= int(temp[4:6]) <= 9):
temp3 = temp[5:6]
else:
temp3 = temp[4:6]
if (1 <= int(temp[6:8]) <= 9):
temp4 = temp[7:8]
else:
temp4 = temp[6:8]
df['trade_date'][i] = temp[:4] + '-' + temp3 + '-' + temp4
# print(df['trade_date'][i])
temp2 = df['trade_date'][i]
# if (choice == 1):
# print("temp2是:", temp2)
df["index_pct_chg"][i] = sh_a_df.iloc[sh_a_df[sh_a_df.date == temp2].index.tolist()[0]]["pct_chg"]
# if (choice == 1):
# print("index是:",df["index_pct_chg"][i])
if(choice==1):
# print(df)
pass
return df
def main():
iter = 0 # 轮数
effective_company = 0 # 有效的公司数量
AR_LIST = [0 for i in range(before_limit+after_limit)]
stock_pctchg_LIST = [0 for i in range(before_limit + after_limit)]
index_pctchg_LIST = [0 for i in range(before_limit + after_limit)]
for item in df_array:
print("This is iter:",iter)
iter+=1
# print(item)
date = item[3]
date_to = datetime.date(*map(int,date.split('-'))) # 官员的到访时间
date_end = date_to + timedelta(days=after_window) # 官员的到访时间=>after_window
date_before = date_to + timedelta(days=-before_window) # before_window=>官员的到访时间
date_earlist = date_to + timedelta(days=-time_from) # 历史数据开始时间
date_to_str = str(date_to).replace('-', '')
date_end_str = str(date_end).replace('-', '')
date_before_str = str(date_before).replace('-', '')
date_earlist_str = str(date_earlist).replace('-', '')
code = item[1] # 股票代码
try:
# 验证的数据集
date_set = pro.daily(ts_code=code, start_date=date_earlist_str, end_date=date_before_str)
# 添加指数
date_set = match_index(date_set,0)
# print(date_set['pct_chg'])
# 第一个窗口
window_before_set = pro.daily(ts_code=code, end_date=date_to_str, limit=before_limit)
# print(date_to_str)
# print(window_before_set)
# 添加指数
window_before_set = match_index(window_before_set,-1)
#
# 第二个窗口
# print("end_date:",date_end_str)
window_after_set = pro.daily(ts_code=code, start_date=date_to_str, end_date=date_end_str).iloc[::-1]
window_after_set = (window_after_set[1:after_limit+1]).reset_index()
new_df = window_after_set
# 添加指数
window_after_set = match_index(new_df,1)
# print(window_after_set)
# 用来预测两个窗口的值:
window_before_set_predict,window_after_set_predict = CPAM(str(date_earlist),str(date_before),str(date_to),str(date_end),date_set,window_before_set,window_after_set)
ar_df = pd.concat([window_before_set_predict,window_after_set_predict])
ar_df["ar"] = ar_df['pct_chg'] - ar_df['predict']
# print(ar_df)
ar_df = ar_df.sort_values("number")
# print(ar_df)
# 新增的
# 获取CAR0 CAR[-1,1],CAR[-5,5],CAR[-10,10],CAR[-15,15],CAR[-20,20]
try: # 要小于before_limit和after_limit
CAR_0 = np.array(ar_df.loc[ar_df['number'] == 0])[0][-1]
# print(CAR_0)
# CAR_1
CAR_1 = 0
for i in range(-1,1+1,1):
CAR_1+= np.array(ar_df.loc[ar_df['number'] == i])[0][-1]
CAR_5 = 0
for i in range(-5, 5 + 1, 1):
CAR_5 += np.array(ar_df.loc[ar_df['number'] == i])[0][-1]
CAR_10 = 0
for i in range(-10, 10 + 1, 1):
CAR_10 += np.array(ar_df.loc[ar_df['number'] == i])[0][-1]
CAR_15 = 0
for i in range(-15, 15 + 1, 1):
CAR_15 += np.array(ar_df.loc[ar_df['number'] == i])[0][-1]
CAR_20 = 0
for i in range(-20, 20 + 1, 1):
CAR_20 += np.array(ar_df.loc[ar_df['number'] == i])[0][-1]
# print(CAR_0,CAR_20)
except:
pass
##
# print(ar_df.iloc[:,11])
if(len(ar_df)==len(AR_LIST)):
for i in range(before_limit + after_limit):
AR_LIST[i] += ar_df.iloc[i, -1]
stock_pctchg_LIST[i] += ar_df.iloc[i, 8]
index_pctchg_LIST[i] += ar_df.iloc[i, 11]
# AR_LIST += ar_df["ar"] # 加到AR_LIST里面去
# print(AR_LIST)
effective_company += 1
else:
pass
# 新增: 存入CSV中
company_class = np.array(df_company_class.loc[df_company_class['股票代码'] == code])[0][2]
market_value_total = np.array(df_company_class.loc[df_company_class['股票代码'] == code])[0][0]
reg_capital = np.array(df_company_class.loc[df_company_class['股票代码'] == code])[0][5]
setup_date = np.array(df_company_class.loc[df_company_class['股票代码'] == code])[0][6]
province = np.array(df_company_class.loc[df_company_class['股票代码'] == code])[0][7]
employees = np.array(df_company_class.loc[df_company_class['股票代码'] == code])[0][8]
index_pct_change = np.array(ar_df.loc[ar_df['number'] == 0])[0][-5]
# print(index_pct_change)
try:
# print("111")
if effective_company == 1:
data = [[code, item[2], company_class, market_value_total, item[3], item[4], item[5],reg_capital,setup_date,province,employees,CAR_0,CAR_1,CAR_5,CAR_10,CAR_15,CAR_20,index_pct_change]]
car_df = pd.DataFrame(data, columns=['Stock code', 'Stock Name', 'Industry', 'Total Market Value', 'Date', 'Governer Class', 'Governer Name',
'reg_capital', 'setup_date', 'province', 'employees_num',
'CAR0','CAR[-1,1]','CAR[-5,5]','CAR[-10,10]','CAR[-15,15]', 'CAR[-20,20]','index_pct_chg'
]) # 将第一维度数据转为为行,第二维度数据转化为列,即 3 行 2 列,并设置列标签
else:
data = [code, item[2], company_class, market_value_total, item[3], item[4], item[5],reg_capital,setup_date,province,employees,CAR_0,CAR_1,CAR_5,CAR_10,CAR_15,CAR_20,index_pct_change] # 数据
car_df.loc[len(car_df)] = data
print(effective_company)
# print(car_df)
except:
##
pass
except:
pass
## 新增的
car_df.to_csv("CAR_INDEXS_ALL_company.csv") # 把数据存入其他csv里面去
##
return AR_LIST,effective_company,stock_pctchg_LIST,index_pctchg_LIST
def AR_SUM_DRAW(AR_LIST,effective_company):
# print(AR_LIST, effective_company)
AR_LIST = np.array(AR_LIST)
AR_LIST_MEAN = AR_LIST / effective_company
# AR_LIST_MEAN_df = pd.dataframe(AR_LIST_MEAN)
print("有效公司数量:", effective_company)
AR_LIST_MEAN_SUM = get_ar_sum(np.array(AR_LIST_MEAN))
AR_LIST_MEAN_SUM_df = pd.DataFrame(columns=['mean_AR_sum'], data=AR_LIST_MEAN_SUM)
AR_LIST_MEAN_SUM_df.to_csv("AR_mean_sum.csv")
print("AR_均值求和:", AR_LIST_MEAN_SUM)
import matplotlib.pyplot as plt
input_values = AR_LIST_MEAN_SUM
squares = [i for i in range(-before_limit, after_limit, 1)]
plt.plot(squares, input_values, linewidth=2)
# 设置图标标题,并给坐标轴加上标签
plt.title("CAR-Time"+str(time_from)+'-'+str(before_window), fontsize=24)
plt.xlabel("Days from now", fontsize=14)
plt.ylabel("CAR", fontsize=14)
plt.savefig("CAR-Time.png")
plt.show()
if __name__ == "__main__":
AR_LIST, effective_company,stock_pctchg_LIST,index_pctchg_LIST = main()
AR_SUM_DRAW(AR_LIST, effective_company)
stock_pctchg_LIST_mean,index_pctchg_LIST_mean = np.array(stock_pctchg_LIST)/effective_company,np.array(index_pctchg_LIST)/effective_company
print(stock_pctchg_LIST_mean,index_pctchg_LIST_mean)
stock_pctchg_LIST_mean_df = pd.DataFrame(columns=['stock_pctchg_LIST_mean'], data=stock_pctchg_LIST_mean)
index_pctchg_LIST_mean_df = pd.DataFrame(columns=['index_pctchg_LIST_mean'], data=index_pctchg_LIST_mean)
stock_pctchg_LIST_mean_df.to_csv("stock_pctchg_LIST_mean_df.csv")
index_pctchg_LIST_mean_df.to_csv("index_pctchg_LIST_mean_df.csv")