Binary response state space models with Scale Mixture of Normal Links

Observation-driven state space models with scale mixture of normal links are presented for binary time series as a robust alternative to the usual normal setup which is commonly used in the literature. We develop an efficient Markov chain Monte Carlo (MCMC) estimation procedure for the proposed state space models. An application using the (aggregated) Tokyo rainfall data set (Knorr-Held, 1999) is analised.