thmm() fits topological hidden Markov models with
Brownian motion with drift (including fractional Brownian motion),
Ornstein-Uhlenbeck and non-parametric emission models; Baum-Welch and
Viterbi are implemented in C.rthmm(), rbm(), rou(),
rbridge() and rmarkov() simulate data.predict(), plot(),
simulate(), logLik() and coef()
methods for fitted models, and ari() for comparing state
sequences.