Estimate Latent Classes on a Mixture of Continuous and Dichotomous Data

The hybrid model likelihood as described by Ramos-Goñi et al. (2017) is implemented and and embedded in a latent class framework. The package is based on 'flexmix' and among others contains an M-step-driver as described by Leisch (2004) . Users can, for example, estimate latent classes for EQ-5D value sets and address preference heterogeneity. Both uncensored and censored data are supported. Furthermore, heteroscedasticity can be taken into account. It is possible to control for different covariates on the continuous and dichotomous data and start values can differ between the expected latent classes.