The model based recursive partitioning approach of Zeileis, Hothorn, and Hornik (2008) offers a way to partition the feature space in order to detect parameter instabilities in the parametric model of interest by means of a structural change test framework. Similarly to latent class or mixture models, the aim of model based partitioning is to identify groups of subjects for which the parameters of the parametric model differ. However, in model based partitioning the groups are usually not defined by a latent factor, but by combinations of observed covariates, that are searched heuristically. Thus, model based partitioning can offer a heuristic but easy to interpret alternative to latent class – as well as random or mixed effects – models.
