hmmviterbi2.cont: Viterbi algorithm to decode the latent states in...

In ziphsmm: Zero-Inflated Poisson Hidden (Semi-)Markov Models

Description

Viterbi algorithm to decode the latent states in continuous-time hidden Markov models with covariates

Usage

hmmviterbi2.cont(y, M, workparm, zero_init, emit_x = NULL,
  zeroinfl_x = NULL, timeindex, plot = FALSE, xlim = NULL, ylim = NULL,
  ...)

Arguments

y	the observed series to be decoded
M	number of latent states
workparm	a vector of values for working parameters, which is the last element returned from hmmfit() function. This consists the generalized logit of prior probabilities (except for the 1st state), generalized logit of transition probability matrix (except for the 1st column), the logit of nonzero structural zero proportions, and the log poisson means
zero_init	a vector containing structural zero proportions in each state, e.g. set zero_init[i] to be 0 if the i-th state is a regular poisson, and otherwise 1.
emit_x	matrix of covariates for the log poisson means. Default to NULL.
zeroinfl_x	matrix of covariates for the nonzero structural zero proportions. Default to NULL.
timeindex	a vector containing the time points
plot	whether a plot should be returned
xlim	vector specifying the minimum and maximum on the x-axis in the plot. Default to NULL.
ylim	vector specifying the minimum and maximum on the y-axis in the plot. Default to NULL.
...	further arguments to be passed to the plot() function

Value

decoded series of latent states

References

Walter Zucchini, Iain L. MacDonald, Roland Langrock. Hidden Markov Models for Time Series: An Introduction Using R, Second Edition. Chapman & Hall/CRC

Examples

priorparm <- 0
tpmparm <- c(-1,-2)
zeroindex <- c(1,0)
zeroparm <- c(0,-1,1)
emitparm <- c(2,0.5,-0.5,3,0.3,-0.2)
workparm <- c(priorparm,tpmparm,zeroparm,emitparm)
timeindex <- rep(1,1000)
for(i in 2:1000) timeindex[i] <- timeindex[i-1] + sample(1:4,1)

designx <- matrix(rnorm(2000),nrow=1000,ncol=2)
x <- cbind(1,designx) #has to make the additional 1st column of 1 for intercept
result <- hmmsim2.cont(workparm,2,1000,zeroindex,emit_x=designx,
                      zeroinfl_x=designx,timeindex=timeindex)
y <- result$series
state <- result$state

fit2 <-  fasthmmfit.cont(y=y,x=designx,M=2,prior_init=c(0.5,0.5),
  tpm_init=matrix(c(-0.2,0.2,0.1,-0.1),2,2,byrow=TRUE),
  zero_init=0.4,emit_init=c(7,21), timeindex=timeindex,
  hessian=FALSE, method="BFGS", control=list(trace=1))
decode2 <- hmmviterbi2.cont(y,2,fit2$working_parm,c(1,0),
  emit_x=designx, zeroinfl_x=designx,
  timeindex=timeindex,plot=FALSE)

ziphsmm documentation built on May 2, 2019, 6:10 a.m.