Estep of the EM algorithm for fitting von Mises (non) homogeneous Markov switching auto-regressive models.

Description

Forward-backward algorithm called in fit.MSAR.

Usage

1
2
Estep.MSAR.VM(data, theta, smth = FALSE, verbose = FALSE, 
   covar.emis = NULL, covar.trans = NULL)

Arguments

data

array of univariate or multivariate series with dimension T*N.samples*d. T: number of time steps of each sample, N.samples: number of realisations of the same stationary process, d: dimension.

theta

model's parameter; object of class MSAR. See also init.theta.MSAR.

smth

If smth=FALSE, only the forward step is computed for forecasting probabilities. If smth=TRUE, the smoothing probabilities are computed too.

verbose
covar.emis

covariables for emission probabilities.

covar.trans

covariables for transition probabilities

Value

list including

loglik

log likelihood

probS

smoothing probabilities: P(S_t=s|y_0,\cdots,y_T)

probSS

one step smoothing probabilities: P(S_t=s,S_{t+1}|y_0,\cdots,y_T)

Author(s)

Valerie Monbet, valerie.monbet@univ-rennes1.fr

References

Ailliot P., Bessac J., Monbet V., Pene F., (2014) Non-homogeneous hidden Markov-switching models for wind time series. JSPI.

See Also

fit.MSAR.VM, Mstep.hh.MSAR.VM,Estep.MSAR

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.