I am interested in using Kalman Filtering and more advnaced case of Particle Filtering for developing a sequenctioanl state space model . However I cannot see any particulra tutorial for this puprose that sue stan for State Space model. I just found LiBBi packgae LibBi: Frequently asked questions. However I prefer to use Stan instead of LiBBi.
Based on the fundamental concept behind the State space mode, I beleive that Stan can be useful as a solver due to its NUTS appoach for estimating parameters in fully non linear state space model beyond kalman filtering. But not sure if am in true track or stan will be too costly for such process:
There are a lot of great resources out there for state space modeling in Stan, although they are a bit dispersed. I think you’ll likely need to move between several different resources to get the information you need. If those aren’t enough, posting another topic on this forum with the specific problem and reproducible code is a great way to overcome implementation problems.
Here are a few resources I found helpful when starting on SSMs in Stan:
Jeffrey Arnold’s notes do a great job showing code that layers on model features. It’s a bit older and the narrative is sparse, but I found it to be a great starting place. This includes a Kalman filter implementation.
The sinhrks/stan-statespace repo has a Stan implementation of examples from a state-space modeling textbook, which may be closer to the narrative structure you’re looking for.
Thank you so mcuh for your helpful comments. If in future you found more tutorials on advanced kalman filtering and particle filterinh used by Stan, I would apprecae to post here, other peaple will take benefit of it
I’ve posted sometimes about ctsem, which uses Stan to implement non linear continuous (and discrete) time state space models in R, using both extended kalman filtering or direct sampling of states. Some examples here Posts | Charles Driver or search for other stuff :)
