Question

这是我的数据，我从20170101到20170106交易了三个符号并记录每个符号重新获利。

label = ['symbol','date','relized']
list = [['S1','20170101',11111],['S1','20170102',12221],['S1','20170103',33445],['S2','20170102',-32421],['S2','20170103',-12211],['S2','20170104',9331],['S2','20170105',62211],['S2','20170106',92211],['S3','20170102',20444]]

df = pd.DataFrame(list,columns=label)

  symbol      date  relized
0     S1  20170101    11111
1     S1  20170102    12221
2     S1  20170103    33445
3     S2  20170102   -32421
4     S2  20170103   -12211
5     S2  20170104     9331
6     S2  20170105    62211
7     S2  20170106    92211
8     S3  20170102    20444

所以，如果今天是20170103，那么总重新为33445 +（ - 12211）+20444，

我想做的是，我的初始资本（参数名称：C）是500000，我想计算值R，公式是

R=Std(C + TodayTotalRelized,period=3Days)/Average(Std value of nearly 3 Day, period=3Days)

有什么好方法吗？

Answer 1

您可以使用滚动功能。

df.rolling(windows=3).std()

如何使用pandas计算滚动窗口的标准偏差

1 个答案: