mssm

Provides methods to perform parameter estimation and make analysis of multivariate observed outcomes through time which depends on a latent state variable. All methods scale well in the dimension of the observed outcomes at each time point. The package contains an implementation of a Laplace approximation, particle filters like suggested by Lin, Zhang, Cheng, & Chen (2005) doi:10.1198/016214505000000349, and the gradient and observed information matrix approximation suggested by Poyiadjis, Doucet, & Singh (2011) doi:10.1093/biomet/asq062.