Conversely, HMMs extend these concepts to sequential or time-dependent data, positing that observations are driven by an underlying Markov process with unobserved, or hidden, states.

A family of learning models that integrates a cognitive diagnostic model and a higher-order, hidden Markov model in one framework is proposed. This new framework includes covariates to model skill ...