Open ghost opened 6 years ago
Unfortunately KFAS does not directly support these as they need a bit different type of approach. However, for t-distribution I think in Durbin & Koopman (2012) book there is a straightforward suggestion how to handle this case, which I also used in one project ages ago.
edit: Chapter 10.8.3 of the book likely contains the relevant information.
Hi, I would like to know if there is a way to implement a SSModel in which the observation errors follow a heavy-tailed distribution like Student's t, instead of the Gaussian distribution.