hsmmspec {mhsmm}R Documentation

Hidden semi-Markov model specification

Description

Creates a model specification of a hidden semi-Markov model.

Usage

hsmmspec(init,transition,emission,sojourn,f,r=NULL,mstep=NULL)

Arguments

init Distribution of states at t=1 ie. P(S=s) at t=1
transition The transition matrix of the embedded Markov chain (diagonal must be 0)
emission A list containing the parameters of the emission distribution
sojourn A list containining the parameters and type of sojourn distribtuion (see Details)
f Density function of the emission distribution
r The function used to generate observations from the emission distribution
mstep Re-estimates the parameters of density function on each iteration

Details

The sojourn argument provides a list containing the parameters for the available sojourn distributions. Available sojourn distributions are shifted Poisson, Gamma and non-parametric.

In the case of the Gamma distribution, sojourn is a list with vectors shape and scale (the Gamma parameters in dgamma), both of length J. Where J is the number of states. See reprocows for an example using Gamma sojourn distributions.

In the case of shifted Poisson, sojourn is list with vectors shift and lambda, both of length J. See hsmmfit for an example using shifted Poisson sojourn distributions.

In the case of non-parametric, sojourn is a list containing a M x J matrix. Where entry (i,j) is the probability of a sojourn of length i in state j. See hsmmfit for an example using shifted non-parametric sojourn distributions.

Value

An object of class hsmmspec

Author(s)

Jared O'Connell

References

Guedon, Y. (2003), Estimating hidden semi-Markov chains from discrete sequences, Journal of Computational and Graphical Statistics, Volume 12, Number 3, page 604-639 - 2003

See Also

hsmm, simulate.hsmmspec, predict.hsmm


[Package mhsmm version 0.3.1 Index]