Modeling Interleaved Hidden Processes

Hidden Markov models assume that observations in time series data stem from some hidden process that can be compactly represented as a Markov chain. We generalize this model by assuming that the obser