This small collection of functions evaluates models with respect to their cross-validation RMSE. Any type of model fitting function may be evaluated.
Inspect and run main.m to see this in action.
No rigorous experiments were completed. However, we can make two observations. First, models with the true number of parameters almost always fit best. Second, there is cross over in RMSE between models. This means that what we use to assess the performance of a model is a function of the evaluation parameter k (the number of folds).
This was developed on GNU Octave (version 4.0.0) and no tests were run for Matlab.