Cross-sectional Background Sample Clauses

Cross-sectional Background. In the cross-sectional context, the likelihood function is used to compare multiple roots. Further, several information criteria based on penalized version of the like- lihood function have been proposed for assessing the number of components in a mixture model. A subset of these information criteria relies on what is known as the classification likelihood, LC (ψ). For example, ICL-BIC [3], which was found to perform well in cross-sectional simulation studies [54], is based on the conditional expectation of the classification likelihood. In the EM framework used for estimation of the cross-sectional finite mixture model, the classification likelihood is often referred to as the complete-data likelihood. Recall that the classification or complete log-likelihood for ψ is given by ∑ ∑