Seemingly unrelated linear regression models are introduced in which thedistribution of the errors is a finite mixture of Gaussian components.Identifiability conditions are provided. The score vector and the Hessianmatrix are derived. Parameter estimation is performed using the maximumlikelihood method and an Expectation-Maximisation algorithm is developed. Theusefulness of the proposed methods and a numerical evaluation of theirproperties are illustrated through the analysis of a real dataset.
展开▼
