ABSTRACT
This chapter deals with versions of the latent Markov (LM) model which are formulated by incorporating constraints corresponding to certain hypotheses of interest. These constraints may concern the measurement model or the latent model. In the first case, they are posed on the distribution of the response variables given the latent process, whereas in the second case, they are posed on the distribution of the latent process. Obviously, depending on the application of interest, an LM model may be also formulated through constraints on both components.
