Modelling Unobserved Heterogeneity Using Mixtures

Austrian Science Fund

Finite mixture models are an important method for modelling latent variables and unobserved heterogeneity and the areas of application are widely spread. In the course of this project the statistical characteristics of several extensions of finite mixture models will be investigated by analyzing specification, identification and estimation of these models. In addition the developed methods will be provided in an open-source implementation as extension of R, an environment for statistical computing and graphics. The methodological part of the project will be complemented by applications focusing on finance and marketing.