HiddenSemiMarkov is an R package for specifying hidden semi-Markov models, simulating sequences, fitting models to specific observation sequences and predicting the sequence of hidden states from observation sequences.
It is especially designed to model multivariate sequences with frequent missing data. The probabilities of missing data may be state- or variable-dependent and specified manually or learned from observation sequences.
To install the package from github:
devtools library is installed or install it with:HiddenSemiMarkov package with the command:devtools::install_github("lasy/HiddenSemiMarkov")
or, if you wish to build the vignette for this package:
devtools::install_github("lasy/HiddenSemiMarkov", build_vignettes = TRUE)
Type
and find examples in the vignette by typing browseVignettes(package = "HiddenSemiMarkov") or start specifying a model with the specify_hsmm function (type ?specify_hsmm for documentation).