如何使用 statsmodels 0.14.0 进行 plot ARIMA 预测/预测

Question

In older versions of the python package statsmodels, there was a plot_predict method in the ARIMAResults class. With the new version, there are only forecast , get_forecast , predict and get_prediction methods, but as far as I could see, no built-in plotting method for the results.

Is there a way to get a similar plot to the one the old plot_predict method generated?

Answer 1

This code works with statsmodels 0.14.0:

from statsmodels.graphics.tsaplots import plot_predict

fig, ax = plt.subplots(figsize=(10,8))

ax.plot(pd.DataFrame(train).loc[354:].reset_index(drop=True), '-g', label='observed')

plot_predict(model_fit, start=354, end=365, ax=ax);

Hint: look up source code and search for plot_predict:

source code for statsmodels.graphics.tsaplots