AbstractHMM subtyping

[nantonel]

Hi,

thanks for this nice package!

I was wandering what is the advantage of having the subtype VariateForm AbstractHMM{F<:VariateForm}?

I feel there is no real necessity of having this since you can get it from the vector of distribution.

[nantonel]

I see this is used here for example:

function likelihoods(hmm::AbstractHMM{Univariate}, observations)
    hcat(map(d -> pdf.(d, observations), hmm.D)...)
end

function likelihoods(hmm::AbstractHMM{Multivariate}, observations)
    # OPTIMIZE ?
    ls = zeros(size(observations)[1], size(hmm)[1])
    @inbounds for i = 1:size(hmm)[1], t = 1:size(observations)[1]
        ls[t,i] = pdf(hmm.D[i], view(observations,t,:))
    end
    ls
end

but this different implementation would not be necessary if observations coming from multivariate distributions were Arrays of Arrays.