学习记录653@python时间序列自相关检验之混成检验

原理

代码

import tushare as ts
ts.set_token('你的code')
df=ts.get_k_data('399300', index=True,start='2016-01-01', end='2016-12-31')
df.head()

本接口即将停止更新，请尽快使用Pro版接口：https://tushare.pro/document/2

	date	open	close	high	low	volume	code
0	2016-01-04	3725.86	3470.41	3726.25	3469.01	115370674.0	sz399300
1	2016-01-05	3382.18	3478.78	3518.22	3377.28	162116984.0	sz399300
2	2016-01-06	3482.41	3539.81	3543.74	3468.47	145966144.0	sz399300
3	2016-01-07	3481.15	3294.38	3481.15	3284.74	44102641.0	sz399300
4	2016-01-08	3371.87	3361.56	3418.85	3237.93	185959451.0	sz399300

import numpy as np
# 基于价格的原始收益率
df['r']=(df['close'] - df['close'].shift(1)) / df['close'].shift(1)

# df['close'].shift(1) 下移一行
# 对数收益率定义为ln(e/s),其中e为下一期价格，s为上一期价格
df['rtn']=np.log(df['close'] / df['close'].shift(1))
df=df.dropna()
df.head()

	date	open	close	high	low	volume	code	r	rtn
1	2016-01-05	3382.18	3478.78	3518.22	3377.28	162116984.0	sz399300	0.002412	0.002409
2	2016-01-06	3482.41	3539.81	3543.74	3468.47	145966144.0	sz399300	0.017544	0.017391
3	2016-01-07	3481.15	3294.38	3481.15	3284.74	44102641.0	sz399300	-0.069334	-0.071855
4	2016-01-08	3371.87	3361.56	3418.85	3237.93	185959451.0	sz399300	0.020392	0.020187
5	2016-01-11	3303.12	3192.45	3342.48	3192.45	174638387.0	sz399300	-0.050307	-0.051617

import matplotlib.pyplot as plt
from statsmodels.graphics.tsaplots import plot_acf
%matplotlib inline
fig = plt.figure(figsize=(10,5))
ax1=fig.add_subplot(111)
fig = plot_acf(df['rtn'],ax=ax1,lags=50) #最高50阶
plt.show()

# 对最高十阶的自相关系数进行检验
import statsmodels.api as sm
import pandas as pd
# 检验10个自相关系数
m = 10
acf,q,p = sm.tsa.acf(df['rtn'],nlags=m,qstat=True)
out = np.c_[range(1,11), acf[1:], q, p]
output=pd.DataFrame(out, columns=['lag', "AC", "Q", "P-value"])
output = output.set_index('lag')
output

	AC	Q	P-value
lag
1.0	-0.178514	7.839777	0.005111
2.0	0.156326	13.876707	0.000970
3.0	-0.099271	16.321298	0.000974
4.0	0.056385	17.113263	0.001837
5.0	-0.166990	24.088764	0.000209
6.0	-0.026627	24.266859	0.000466
7.0	0.002090	24.267962	0.001022
8.0	0.000263	24.267979	0.002066
9.0	0.116595	27.726689	0.001059
10.0	-0.006233	27.736615	0.001989

以0.05为置信水平的话，这些p值均小于0.05，因此拒绝原假设，得出结论：自相关是显著的，可信的。