A Finite Mixture Kullback - Leibler Model Selection Criterion

A Finite Mixture Kullback - Leibler Model Selection Criterion

The purpose of statistical modeling is to construct a model that approximates the tre structure as accurately as possible through the use of available data. A good model will generally yield good results; however, one cannot expect to obtain good results when using an inappropriate model. Herein lies the importance of model evaluation criteria for assessing the goodness of a subjective model. This paper considers a finite mixture of the known Kullback-Leibler criterion to the model selection problem. The aim of this criterion is to select an admissible set of models based on a measure of closeness. We demonstrate that a very general class of statistical criterion, which we call that finite mixture Kullback-Leibler criterion, provides a way of rival theory model selection