持有时间与仓位算法
- 读取本地数据
- 读取信号, 并且修改权重
- 持有10天的权重分配
1_读取本地数据
import pandas as pdimport numpy as npfrom jaqs.data import DataViewfrom jaqs.research import SignalDiggerimport warningsdata = pd.read_excel('signal_df.xlsx', index_col='trade_date')
print(data.tail().stack())
trade_date
20171218 000001.SZ 0.028519
000069.SZ 0.044093
000100.SZ 0.028474
000402.SZ 0.029721
000623.SZ 0.037530
000728.SZ 0.045084
000776.SZ 0.029339
000783.SZ 0.037970
000959.SZ 0.027329
000963.SZ 0.046964
002236.SZ 0.106231
002241.SZ 0.026804
002310.SZ 0.028444
002426.SZ 0.026606
002456.SZ 0.160243
002475.SZ 0.062133
002572.SZ 0.064084
002736.SZ 0.030157
002831.SZ 0.026710
002839.SZ 0.045822
300059.SZ 0.040967
600016.SH 0.028757
600018.SH 0.036246
600048.SH 0.029588
600068.SH 0.066481
600297.SH 0.029303
600340.SH 0.032477
600369.SH 0.032991
600376.SH 0.044635
600383.SH 0.042118
...
20171222 600376.SH 0.069270
600390.SH 0.026247
600415.SH 0.046980
600522.SH 0.029764
600663.SH 0.085307
600795.SH 0.052296
600804.SH 0.135629
600827.SH 0.051967
600837.SH 0.071733
600871.SH 0.036321
600895.SH 0.048910
600926.SH 0.037022
600958.SH 0.066338
600959.SH 0.053739
601018.SH 0.035732
601118.SH 0.031308
601186.SH 0.025082
601198.SH 0.027028
601212.SH 0.066140
601228.SH 0.036832
601229.SH 0.025712
601555.SH 0.042266
601633.SH 0.025228
601727.SH 0.030271
601985.SH 0.025059
601992.SH 0.034200
601997.SH 0.045971
603160.SH 0.037208
603799.SH 0.065817
603993.SH 0.034091
Length: 283, dtype: float64
2_读取信号, 并且修改权重
import pandas as pdweight_list = []for time_index, weight in data.iterrows(): weiht_result = (weight/weight.sum()) weight_list.append(weiht_result.to_dict())stock_df = pd.DataFrame(weight_list, index=data.index)
print(stock_df.tail().stack())
trade_date
20171218 000001.SZ 0.012165
000069.SZ 0.018808
000100.SZ 0.012145
000402.SZ 0.012678
000623.SZ 0.016008
000728.SZ 0.019231
000776.SZ 0.012515
000783.SZ 0.016196
000959.SZ 0.011657
000963.SZ 0.020032
002236.SZ 0.045312
002241.SZ 0.011433
002310.SZ 0.012133
002426.SZ 0.011349
002456.SZ 0.068351
002475.SZ 0.026503
002572.SZ 0.027335
002736.SZ 0.012863
002831.SZ 0.011393
002839.SZ 0.019545
300059.SZ 0.017474
600016.SH 0.012266
600018.SH 0.015460
600048.SH 0.012621
600068.SH 0.028357
600297.SH 0.012499
600340.SH 0.013853
600369.SH 0.014072
600376.SH 0.019039
600383.SH 0.017965
...
20171222 600376.SH 0.025168
600390.SH 0.009537
600415.SH 0.017069
600522.SH 0.010815
600663.SH 0.030995
600795.SH 0.019001
600804.SH 0.049279
600827.SH 0.018882
600837.SH 0.026063
600871.SH 0.013197
600895.SH 0.017771
600926.SH 0.013452
600958.SH 0.024103
600959.SH 0.019525
601018.SH 0.012983
601118.SH 0.011375
601186.SH 0.009113
601198.SH 0.009820
601212.SH 0.024031
601228.SH 0.013382
601229.SH 0.009342
601555.SH 0.015357
601633.SH 0.009166
601727.SH 0.010998
601985.SH 0.009105
601992.SH 0.012426
601997.SH 0.016703
603160.SH 0.013519
603799.SH 0.023914
603993.SH 0.012387
Length: 283, dtype: float64
print(stock_df.sum(axis=1))
trade_date
20120105 1.0
20120106 1.0
20120109 1.0
20120110 1.0
20120111 1.0
20120112 1.0
20120113 1.0
20120116 1.0
20120117 1.0
20120118 1.0
20120119 1.0
20120120 1.0
20120130 1.0
20120131 1.0
20120201 1.0
20120202 1.0
20120203 1.0
20120206 1.0
20120207 1.0
20120208 1.0
20120209 1.0
20120210 1.0
20120213 1.0
20120214 1.0
20120215 1.0
20120216 1.0
20120217 1.0
20120220 1.0
20120221 1.0
20120222 1.0
...
20171113 1.0
20171114 1.0
20171115 1.0
20171116 1.0
20171117 1.0
20171120 1.0
20171121 1.0
20171122 1.0
20171123 1.0
20171124 1.0
20171127 1.0
20171128 1.0
20171129 1.0
20171130 1.0
20171201 1.0
20171204 1.0
20171205 1.0
20171206 1.0
20171207 1.0
20171208 1.0
20171211 1.0
20171212 1.0
20171213 1.0
20171214 1.0
20171215 1.0
20171218 1.0
20171219 1.0
20171220 1.0
20171221 1.0
20171222 1.0
Length: 1452, dtype: float64
3_持有gap天的权重分配
- 切片拆分
- 重塑索引填补权重
- 将表格求和
import pandas as pdimport numpy as npfrom datetime import datetime, timedeltadef cal_daily_weight(df, gap=5): df = pd.concat([pd.DataFrame(0, list(range(gap-1)), df.columns), df]) return (pd.rolling_apply(df, gap, sum)/gap).iloc[gap-1:]def change_columns_index(signal): new_names = {} for c in signal.columns: if c.endswith('SZ'): new_names[c] = c.replace('SZ', 'XSHE') elif c.endswith('SH'): new_names[c] = c.replace('SH', 'XSHG') signal = signal.rename_axis(new_names, axis=1) signal.index = pd.Index(map(lambda x: datetime.strptime(str(x),"%Y%m%d") , signal.index)) signal.index = pd.Index(map(lambda x: x+timedelta(hours=15) , signal.index)) return signaltable = change_columns_index(stock_df)target = cal_daily_weight(table.fillna(0), gap=8)print(target.head(9).sum(1))
D:\Anaconda3\lib\site-packages\ipykernel_launcher.py:7: FutureWarning: pd.rolling_apply is deprecated for DataFrame and will be removed in a future version, replace with
DataFrame.rolling(window=8,center=False).apply(func=<builtin_function_or_method>,args=<tuple>,kwargs=<dict>)
import sys
2012-01-05 15:00:00 0.125
2012-01-06 15:00:00 0.250
2012-01-09 15:00:00 0.375
2012-01-10 15:00:00 0.500
2012-01-11 15:00:00 0.625
2012-01-12 15:00:00 0.750
2012-01-13 15:00:00 0.875
2012-01-16 15:00:00 1.000
2012-01-17 15:00:00 1.000
dtype: float64
