Clustering Based on Multivariate Mixed Type Longitudinal Data with an application to the EU-SILC database

Abstract

We present a statistical model for joint modelling of several mixed-type longitudinal outcomes, while performing unsupervised clustering with respect to different patterns. Method is demonstrated on the EU-SILC dataset consisting of Czech households followed in a time span 2005 - 2018.

Keywords

• Multivariate longitudinal data
• Mixed type outcome
• Model based clustering
• EU-SILC