search

lindermanlab / ssm

Bayesian learning and inference for state space models
MIT License
540 stars 196 forks source link

Constraining C during PLDS model fitting #135

Closed erinshappell closed 1 year ago

erinshappell commented 2 years ago

I'm interested in fitting a PLDS model with partial constraints on C--is there a way to achieve this using SSM? If not, could this be achieved by setting priors on the observation matrix (and if so, where should I look to set those priors)? Thanks!

slinderman commented 2 years ago

Hi Erin, that's certainly possible but it would involve a little code modification. Do you want to fix the other emissions parameters (i.e. the input weights, bias, and covariance matrix) too? If so, a simple thing to do is set the emissions model m_step to a no-op:

slds.emissions.m_step = lambda *args, **kwargs: None