Can I fix some parameter in EstimateProcess
?
I want to estimate the transition matrix of a hidden markov process without optimization/change of the distribution for the states
Example:
hmmProcess =
HiddenMarkovProcess[{t0, t1}, {{a11, a12}, {a21, a22}}, {NormalDistribution[0.2, 0.1],
NormalDistribution[0.85, 0.1]}]
hmm = EstimatedProcess[trace, hmmProcess, {{a11, 0.9}, {a12, 0.1}, {a21, 0.1}, {a22, 0.9}, {t0, 1}, {t1, 1.0*^-20}}]
The output is still a process with changed/optimized state distribution:
HiddenMarkovProcess[{1., 6.19154*10^-104}, {{0.953574, 0.046426}, {0.0318546, 0.968145}}, {NormalDistribution[0.202469, 0.0807677], NormalDistribution[0.732095, 0.0819761]}]