(1) Can cim in Equation (6) be continuous variable which indicating that degree of involvement of component in item? Just as qimk can be binary and continuous variable.
(2) In section 5, it was said 'component item difficulties should be easy to moderate when multiple components are involved' such that sufficient information to recover the model parameters. I'm really wondering the reason.
(3) The comparison between DINA with full MLTM-D, and LCDM with full-MLTM-D is award. Why not consistently use MCMC or ML with EM to estimate the models?