Heterogeneity in many datasets stems from the different behaviors of several underlying groups or subpopulations. The aim of this paper is to classify observations in such a dataset into these latent groups when each group's behavior is piecewise linearly related to a set of covariates. We...