decodeHSMM: State decoding

In PHSMM: Penalised Maximum Likelihood Estimation for Hidden Semi-Markov Models

Description

State decoding for the HSMM estimated using pmleHSMM. Decoding is based on the Viterbi algorithm and the corresponding HMM model representation.

Usage

decodeHSMM(y, mod)

Arguments

y	vector containing the observed time series.
mod	model object as returned by pmleHSMM.

Value

Returns a vector containing the decoded states.

References

For more details about the Viterbi algorithm, see for example:

Zucchini, W., MacDonald, I.L. and Langrock, R. (2016): Hidden Markov models for time series: An introduction using R. 2nd edition. Chapman & Hall/CRC, Boca Raton.

Examples

# running this example might take a few minutes
#
# 1.) 2-state gamma-HSMM for hourly muskox step length
# with an unstructured start of length of 10
#
# initial values
p_list0<-list()
p_list0[[1]]<-c(dgeom(0:9,0.2),1-pgeom(9,0.2))
p_list0[[2]]<-c(dgeom(0:9,0.2),1-pgeom(9,0.2))
mu0<-c(5,150)
sigma0<-c(3,180)
#
# fit 2-state gamma-HSMM with lambda=c(100,100)
# and difference order 3
# estimation might take a few minutes
PHSMM<-pmleHSMM(y=muskox$step,N=2,p_list=p_list0,mu=mu0,
                sigma=sigma0,lambda=c(100,100),order_diff=3,
                y_dist='gamma')
#
# state decoding
s_HSMM<-decodeHSMM(muskox$step,mod=PHSMM)
# plot sequence of the decoded time series
plot(muskox$step[1:1000],type='h',xlab='time (h)',ylab='step (m)',
  main='',col=s_HSMM)
legend('topright',c('state 1','state 2'),lwd=2,col=1:2)




# running this example might take a few minutes
#
# 2.) 3-state gamma-HSMM for hourly muskox step length
# with an unstructured start of length of 10
#
# initial values
p_list0<-list()
p_list0[[1]]<-c(dgeom(0:9,0.2),1-pgeom(9,0.2))
p_list0[[2]]<-c(dgeom(0:9,0.2),1-pgeom(9,0.2))
p_list0[[3]]<-c(dgeom(0:9,0.2),1-pgeom(9,0.2))
omega0<-matrix(0.5,3,3)
diag(omega0)<-0
mu0<-c(5,100,350)
sigma0<-c(3,90,300)
#
# fit 3-state gamma-HSMM with lambda=c(1000,1000,1000)
# and difference order 3
# estimation might take some minutes
PHSMM<-pmleHSMM(y=muskox$step,N=3,p_list=p_list0,mu=mu0,
                sigma=sigma0,omega=omega0,
                lambda=c(1000,1000,1000),
                order_diff=3,y_dist='gamma')
#
# state decoding
s_HSMM<-decodeHSMM(muskox$step,mod=PHSMM)
# plot sequence of the decoded time series
plot(muskox$step[1:1000],type='h',xlab='time (h)',ylab='step (m)',
  main='',col=s_HSMM)
legend('topright',c('state 1','state 2', 'state 3'),lwd=2,col=1:3)

PHSMM documentation built on Feb. 9, 2021, 5:07 p.m.