从ARIMA提取摘要值

Question

运行ARIMA模型后，我面临以下问题：

model_final=ARIMA(data_set_final["Price_DX"], order = (ar_order,0,ma_order), exog = data_set_exog)

    SARIMAX Results
==============================================================================
Dep. Variable:           Price_DX   No. Observations:                   42
Model:                 ARIMA(1, 0, 0)   Log Likelihood                -156.392
Date:                Mon, 26 Jul 2021   AIC                            322.784
Time:                        20:48:33   BIC                            331.472
Sample:                    07-01-2010   HQIC                           325.968
                         - 10-01-2020
Covariance Type:                  opg
==================================================================================
                     coef    std err          z      P>|z|      [0.025      0.975]
----------------------------------------------------------------------------------
const           -101.4037     57.505     -1.763      0.078    -214.112      11.304
Price_DX1     0.1354      0.053      2.554      0.011       0.032       0.239
Europe_DX1      1.1445      0.647      1.768      0.077      -0.124       2.413
ar.L1              0.4449      0.164      2.718      0.007       0.124       0.766
sigma2            99.8929     26.295      3.799      0.000      48.356     151.430
===================================================================================
Ljung-B`enter code here`ox (L1) (Q):                   0.60   Jarque-Bera (JB):                 0.02
Prob(Q):                              0.44   Prob(JB):                         0.99
Heteroskedasticity (H):               0.68   Skew:                            -0.04
Prob(H) (two-sided):                  0.49   Kurtosis:                         3.06
===================================================================================

如何从ARIMA摘要表中提取Prob(Q)和Prob(H)值？

例如，我可以通过键入以下内容轻松获得AIC：

print(model_final_fit.aic)

不幸的是，我找不到Ljung-Box和异方差性这里的属性。你知道怎么容易弄到吗？

Answer 1

摘要方法将这些输出存储为html表。您可以通过转换成熊猫数据来提取这些值。

test = pd.read_html(model_final.summary().tables[2].as_html(),header=None,index_col=0)[0]
# Prob(Q) 
print(test[1].iloc[1])

#Prob(H)
print(test[1].iloc[3])