simulate.hmm: Simulating Hidden Markov objects

In asbjornholk/HiddenStateModels: Fit or Create Hidden State Models

Description

Simulating Hidden Markov objects

Usage

## S3 method for class 'hmm'
simulate(object, nsim = 1, seed = NULL, include_state = FALSE, ...)

Arguments

object	Object of class 'hmm'.
nsim	Number of simulations.
seed	Seed to be used for random generation.
include_state	Logical, whether or not the hidden state should also be returned.
...	Additional arguments.

Value

If observations are available in the 'hmm' object, will return a matrix with 'nsim' rows, each of which being a simulation of equal length as the observed data. Otherwise, will return a vector of length 'nsim' of simulations from the model. In either case, if 'include_state' is 'TRUE', each simulation will be appended with a vector of equal length indicating the hidden state.

Examples

# Continuing examples from hmm page

# Normal distributions
X.normal <- simulate(hmm.normal, nsim=100)
summary(X.normal)
plot(X.normal)
hist(X.normal)

# Custom (uniform) distributions
Z.unif <- simulate(hmm.unif, nsim=200, include_state=TRUE)
X.unif <- Z.unif[1:100]
G.unif <- Z.unif[101:200]
plot(X.unif, type='h', col=G.unif, lwd=2)

# Custom (mixed) distributions
Z.mixture <- simulate(hmm.mixture, nsim=200, include_state=TRUE)
X.mixture <- Z.mixture[1:200]
G.mixture <- Z.mixture[201:400]
plot(X.mixture, type='h', col=G.mixture, lwd=2)

asbjornholk/HiddenStateModels documentation built on Jan. 9, 2022, 12:02 a.m.