R implementation of HMMs described in processed report except function HMMLikelihood renamed to R_HMMLikelihood and changed to compute values for all capture histories and return lnl, alpha, phi, v, dmat, and gamma values. loglikelihood is called with a fitted hmm model and then computes the gamma,dmat and delta matrices and calls R_HMMLikelihood function. These are not used by the fitting code.
R_HMMLikelihood(x,first,m,T,dmat,gamma,delta)
loglikelihood(object,ddl=NULL)
both return log-likelihood, alpha, v and phi arrays
single observed sequence (capture history)
occasion to initiate likelihood calculation for sequence
number of states
number of occasions; sequence length
observation probability matrices
transition matrices
initial distribution
fitted hmm model
design data list; will be computed if NULL
Jeff Laake
Zucchini, W. and I.L. MacDonald. 2009. Hidden Markov Models for Time Series: An Introduction using R. Chapman and Hall, Boca Raton, FL. 275p. See page 45.