I recently started learning HMM and was wondering how do I go about using a model or similar thereof in which an observation is really a realization of the Gaussian distribution of the corresponding hidden states? (Each hidden state would correspond to a Gaussian distribution with unique mean, standard deviation,…)
I suppose the state transition probability doesn't have to be continuous.
Are there any models based on this? or similar to this?
Best Answer
I don't think that HMM is anything to do with distribution of hidden probabilities. Though distribution of response variable matters. Transition probabilities could come through forward backward algo. if there are no co-variates then fixed matrix would come else it would be functions of PDF for all states( as a function of external variables/co-variates).