This paper1 advances results in model selection by relaxing the task of optimally tuning the regularization parameter in a number of algorithms with respect to the classical cross-validation performance criterion as a convex optimization problem. The proposed strategy differs from the scope of e.g. generalized cross-validation (GCV) as it concerns the efficient optimization, not the individual ...

A problem in model selection, namely the identification of multiple change points for a piece-wise constant hazard rate, is discussed. A methodology using the Bayes’ Information Criterion is developed in an overdispersed survival model (with corresponding quasi-likelihood function). The technique is used to identify changes in the historical frequency of forest fire. It is applied to two datase...

One of the most popular criteria for model selection is the Bayesian Information Criterion (BIC). It is based on an asymptotic approximation using Bayes rule when the sample size tends to infinity and the dimension of the model is fixed. Although it works well in classical applications, it performs less satisfactorily for high dimensional problems, i.e. when the number of regressors is very lar...

We describe a procedure for inducing conditional densities within the mixtures of truncated exponentials (MTE) framework. We analyse possible conditional MTE specifications and propose a model selection scheme, based on the BIC score, for partitioning the domain of the conditioning variables. Finally, experimental results demonstrate the applicability of the learning procedure as well as the ex...