I'm writing an open source software toolkit for multiscale
autoregressive modeling with wavelets (http://www.dataraptor.net/
MSARW/). As part of an upgrade, automatic fitting procedures are being
added. For this upgrade, I need the mean and variance, a way of
testing values of the parameters listed below or both. I've made an
extensive Google search and found nothing. Any help is appreciated.
The parameters in question are:
AIC (Akaike Information Criterion)
http://en.wikipedia.org/wiki/Akaike_information_criterion
BIC (Bayesian Information Criterion)
http://en.wikipedia.org/wiki/Bayesian_information_criterion
MAE (Mean Absolute Error)
http://en.wikipedia.org/wiki/Mean_absolute_error
Thank you