mvnHMM: Multivariate Normal Mixture Model

In SWotherspoon/MVNSeq: Normal Mixture and Hidden Markov Models

Multivariate Normal Mixture Model

Description

Fit multivariate Normal Hidden Markov models by EM.

Usage

mvnHMM(
  y,
  cl,
  min.iters = 10,
  max.iters = 50,
  tol = 0.001,
  verbose = interactive()
)

gmvnHMM(
  y,
  cl,
  gr,
  common.transition = FALSE,
  min.iters = 10,
  max.iters = 50,
  tol = 0.001,
  verbose = interactive()
)

grmvnHMM(
  y,
  cl,
  gr,
  common.transition = FALSE,
  min.iters = 10,
  max.iters = 100,
  tol = 0.001,
  verbose = interactive()
)

Arguments

y	the sequence of observations
cl	an integer vector allocating observations to classes
min.iters	minimum number of EM iterations
max.iters	maximum number of EM iterations
tol	tolerance for the log likelihood
verbose	should the log likelihood be reported.
gr	an integer vector allocating observations to groups
common.transition	should the transition probabilities be common across groups

Details

These functions fit K state multivariate Normal hidden Markov models to sequences of observations. 'mvnHMM' fits a hidden Markov model to a single sequence of observations. 'gmvnHMM' and 'grmvnHMM' fits separate Markov models to several groups (sequences) of observations. 'gmvnHMM' fits a hidden Markov model to each group so that each state has a different mean across groups, but a common covariance. 'grmvnHMM' constrains the means of the states to be Normally distributed across groups.

Value

the fitted model

SWotherspoon/MVNSeq documentation built on June 10, 2025, 6:15 a.m.