search

lindermanlab / ssm

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

access lag-specific autoregressive A matrix #148

Closed r-shruthi11 closed 2 years ago

r-shruthi11 commented 2 years ago

For a specific lag hmm.observations.As[K, :, l * D : (l + 1) * D] corresponds to the weights that map onto lth lagged data as the definition here suggests so if I want to access the coefficients for the linear regression one frame back for state k this would be hmm.observations.As[k, :, 0 : D]? Or are these the coefficients for the hmm.observations.lags - l frame? I'm a little confused by this block here and how similiar l and 1 look!

r-shruthi11 commented 2 years ago

Did some tests and confirming that indeed indexing into hmm.observations.As works per the definition here

Closing this issue.