Model selection: Cp, AIC, BIC and adjusted R²

Significance and meaning of Cp, AIC, BIC and adjusted R²

While working on time series analysis project that weather recession will hit and when it will hit I came across to use statsmodels library which is similar to sklean I used their ARIMA and SARIMA models. When I used it there is one step comes in which they gives summary of model and there are so many different values which is the title of this article.

This is how ARIMA model statistics look like:

And you can spot AIC and BIC values in this summary table.

I tried to read and learn online about AIC, BIC and Cp but there is no satisfactory or I would say simple explanation to it. So I thought I should write article on it.

For simplicity I am considering an example which is linear regression least square fit model.

What these values means for any machine learning model and how they are going to make any help in making decisions?

For any model we are always more focused on test error rather than train error. Because there are so many times possible that your…