Parameter estimation (via Nelder-Mead / Optim.jl) should be working properly now, although ultimately I think we should be looking at implementing EM estimation (particularly now that we have the beginnings of missing data support, and with Nullable on the release horizon)?
Addresses https://github.com/JuliaStats/TimeModels.jl/issues/34 and the current failing test issue, adds some new tests, and cleans things up a bit.
Parameter estimation (via Nelder-Mead / Optim.jl) should be working properly now, although ultimately I think we should be looking at implementing EM estimation (particularly now that we have the beginnings of missing data support, and with
Nullable
on the release horizon)?