Looking for suitable methods to check goodness of a model based on the residuals or prediction errors?

Question

I am looking for suitable methods to check goodness of my model based on the residuals or prediction errors. I want to make sure that the model residuals or prediction errors are normally distributed and uncorrelated. I would appreciate your suggestions. Thanks, Iman

Answer 1

ACF(yourmodel$residuals) and ACF(yourmodel$residuals^2): if both correlograms suggest that the residuals of the fitted model behave like white noise, you can make sure you did right.