Validation, or more specifically, cross-validation is a method of assessing how well
the regression can be generalized.
One approach it to retain a portion of the original data for validation purposes and use
the rest to construct the model. Once the model is finalized, the estimated parameters are used to
predict the out-of-sample data.
The model’s validity may be assessed by comparing the out-of-sample mean squared error
(the mean squared prediction error), with the in-sample mean square error. If the out-of-sample mean
squared error is substantially higher, that would imply deficiency in the model’s ability to predict.
