Simulate data from an autoregressive HMM — sample

Simulate data from an autoregressive HMM with AR(p). Different distributions can be specified in dists (uni- and multivariate).

Usage

sample_arp(n_samples, delta, Gamma, N, params, autocor, p, dists)

n_samples: Number of samples to generate.
delta: Initial distribution of the Markov chain.
Gamma: Transition probability matrix of the Markov chain.
N: Number of states.
params: Parameter vector for the different distributions. . Has to respect the order specified in dists.
autocor: List of lists of parameters for the autoregression parameters. Has to match p, in the order \(\phi_{t-p},\dots,\phi_{t-1}\) where \(\phi\) is the vector of autoregression parameters for one specific time lag (one value for each state, some values can be NA). 0, if no autoregression. Has to respect the order specified in dists.
p: Vector of degree of autoregression for each distribution, 0 = no autoregression. Allows individual values for every state (one entry for every state of every variable)
dists: Vector containing abbreviated names (in R-jargon) of the distributions to be considered in the likelihood computation.

List of states and data of the HMM.