Julia based implementation of a latent Markov model for analysing questionnaire data. WIP The present implementation is merely meant to play around with Turing.jl for obtaining an efficient algorithm for a model similar to that proposed in
Bartolucci, Lupparelli and Montanari Latent Markov model for longitudinal binary data: An application to the performance evaluation of nursing homes The Annals of Applied Statistics, 611--636, 2009.
In this implementation, the likelihood is computed from the backward information filter and latent states are marginalised out. Frequentist estimates (based on the EM-algorithm) can be obtained from the LMest package in R.