Curve Clustering with Random Effects Regression Mixtures

Abstract

In this paper we address the problem of clustering sets of curve or trajectory data generated by groups of objects or individuals. The focus is to model curve data directly using a set of model-based curve clustering algorithms referred to as mixtures of regressions or regression mixtures. The proposed methodology is based on extension to regression mixtures that we call random effects regression mixtures which combines linear random effects models with standard regression mixtures. We develop a general expectationmaximization (EM) algorithm using maximum a posteriori (MAP) estimation for random effects regression mixtures and demonstrate how this technique can be applied to the problem of clustering cyclone data.