hhsmmspec function

hhsmm specification

Specify a model of class "hhsmmspec" using the model parameters

hhsmmspec(
  init,
  transition,
  parms.emission,
  sojourn = NULL,
  dens.emission,
  remission = NULL,
  mstep = NULL,
  semi = NULL
)

Arguments

• init: vector of initial probabilities

• transition: the transition matrix

• parms.emission: the parameters of the emission distribution

• sojourn: the sojourn distribution, which is one of the following cases:

  • a list containing d, which is a nobs

    (number of observations) times nstates (number of states) matrix of probabilities, and type = "nonparametric" for non-parametric sojourn distribution

  • a list containing the parameters mu, shift and size of a shifted negative binomial distribution, for each semi-Markovian state, and type = "nbinom" for negative binomial sojourn distribution

  • a list containing the parameters shape and shift of a shifted logarithmic distribution, for each semi-Markovian state, and type = "logarithmic" for logarithmic sojourn distribution

  • a list containing the parameters lambda and shift of the shifted poisson distribution, for each semi-Markovian state, and type = "poisson" for Poisson sojourn distribution

  • a list containing the parameters shape and scale of the gamma distribution, for each semi-Markovian state, and type = "gamma" for gamma sojourn distribution

  • a list containing the parameters shape and scale of the Weibull distribution, for each semi-Markovian state, and type = "weibull" for Weibull sojourn distribution

  • a list containing the parameters meanlog and sdlog of the log-normal distribution, for each semi-Markovian state, and type = "lnorm" for log-normal sojourn distribution

• dens.emission: the probability density function of the emission

• remission: the random sample generation from the emission distribution

• mstep: the M step function for the EM algorithm

• semi: a logical vector of length nstate: the TRUE associated states are considered as semi-markov

Returns

a model of class "hhsmmspec"

Examples

init = c(1, 0)
transition = matrix(c(0, 1, 1, 0), 2, 2)
parms.emission = list(mix.p = list(c(0.5, 0.5), 1),
                                mu = list(list(c(1, 2), c(5, 1)), c(2, 7)),
              sigma = list(list(diag(2), 2 * diag(2)), 0.5 * diag(2)))
sojourn = list(lambda = 1, shift = 5, type = "poisson")
dens.emission = dmixmvnorm
remission = rmixmvnorm
mstep = mixmvnorm_mstep
semi = rep(TRUE,2)
model = hhsmmspec(init, transition, parms.emission, sojourn, 
dens.emission, remission, mstep, semi)

Author(s)

Morteza Amini, morteza.amini@ut.ac.ir , Afarin Bayat, aftbayat@gmail.com