Clustering bivariate mixed-type data via the cluster-weighted model