Multioutput Kalman filter with scan and missing observations

Hi all, I’m trying to fit a multivariate linear Gaussian state space model using the Kalman filter. Some observations are missing, so if I want to use a mv-normal sampling statement at each iteration the dimension of the multivariate normal distribution can vary between iterations. I can’t think of a way to do this with all the restrictions Jax puts on Boolean indexing. Any ideas?

I would recommend using tfp.sts module which supports missing outputs.

1 Like