Question

我有一个包含以下值的csv文件：

Date        Value
15/01/2016  0.6972
14/01/2016  0.6985
13/01/2016  0.6955
12/01/2016  0.6986
11/01/2016  0.6995
8/01/2016   0.6953

这是一个超过5年的日常数据的时间序列。如何访问pandas来创建一个ewma（10）（指数跨度为10的指数加权移动平均线？

非常感谢！

Answer 1

IIUC你可以尝试ewma：

import pandas as pd
import numpy as np
import io

temp=u"""Date;Value
15/01/2016;0.6972
14/01/2016;0.6985
13/01/2016;0.6955
12/01/2016;0.6986
11/01/2016;0.6995
8/01/2016;0.6953"""

#change io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), sep=";", index_col=None, parse_dates=[0])
print df

        Date   Value
0 2016-01-15  0.6972
1 2016-01-14  0.6985
2 2016-01-13  0.6955
3 2016-12-01  0.6986
4 2016-11-01  0.6995
5 2016-08-01  0.6953

print pd.ewma(df['Value'], span=10)

0    0.697200
1    0.697915
2    0.696944
3    0.697490
4    0.698067
5    0.697348
Name: Value, dtype: float64

或类似的东西：

import pandas as pd
import io

temp=u"""Date;Value
15/01/2016;0.6972
14/01/2016;0.6985
13/01/2016;0.6955
12/01/2016;0.6986
11/01/2016;0.6995
8/01/2016;0.6953"""

#change io.StringIO(temp) to filename
df = pd.read_csv(io.StringIO(temp), sep=";", index_col=[0], parse_dates=[0])
print df

             Value
Date              
2016-01-15  0.6972
2016-01-14  0.6985
2016-01-13  0.6955
2016-12-01  0.6986
2016-11-01  0.6995
2016-08-01  0.6953

print pd.ewma(df['Value'], span=10, freq='D')

Date
2016-01-13    0.695500
2016-01-14    0.697150
2016-01-15    0.697170
2016-01-16    0.697170
2016-01-17    0.697170
2016-01-18    0.697170
2016-01-19    0.697170
2016-01-20    0.697170
2016-01-21    0.697170
2016-01-22    0.697170
2016-01-23    0.697170
2016-01-24    0.697170
2016-01-25    0.697170
2016-01-26    0.697170
2016-01-27    0.697170
2016-01-28    0.697170
2016-01-29    0.697170
2016-01-30    0.697170
2016-01-31    0.697170
2016-02-01    0.697170
2016-02-02    0.697170
2016-02-03    0.697170
2016-02-04    0.697170
2016-02-05    0.697170
2016-02-06    0.697170
2016-02-07    0.697170
2016-02-08    0.697170
2016-02-09    0.697170
2016-02-10    0.697170
2016-02-11    0.697170
                ...   
2016-11-02    0.699500
2016-11-03    0.699500
2016-11-04    0.699500
2016-11-05    0.699500
2016-11-06    0.699500
2016-11-07    0.699500
2016-11-08    0.699500
2016-11-09    0.699500
2016-11-10    0.699500
2016-11-11    0.699500
2016-11-12    0.699500
2016-11-13    0.699500
2016-11-14    0.699500
2016-11-15    0.699500
2016-11-16    0.699500
2016-11-17    0.699500
2016-11-18    0.699500
2016-11-19    0.699500
2016-11-20    0.699500
2016-11-21    0.699500
2016-11-22    0.699500
2016-11-23    0.699500
2016-11-24    0.699500
2016-11-25    0.699500
2016-11-26    0.699500
2016-11-27    0.699500
2016-11-28    0.699500
2016-11-29    0.699500
2016-11-30    0.699500
2016-12-01    0.698602
Freq: D, Name: Value, dtype: float64

Exponentially weighted moment functions - docs

读取时间序列csv并计算熊猫中的ewma

1 个答案: