HMMHSMM (version 0.1.0)

localdecodeHMM: Local Decoding for Hidden Markov Models

Description

Computes the most likely hidden state sequence for a univariate time series using a fitted Hidden Markov Model (HMM) object via the forward-backward algorithm.

Usage

localdecodeHMM(x, HMM, obsdist)

Value

Numeric vector of length length(x), containing the most likely hidden state at each observation, computed via local decoding using the forward-backward algorithm.

Arguments

x

Numeric vector. The observed data sequence.

HMM

List. A fitted HMM object, typically returned by findmleHMM. Must contain estimated parameters delta and Pi, as well as state-dependent observation parameters.

obsdist

Character string. The observation distribution used in the fitted HMM. Supported distributions: "norm", "pois", "weibull", "zip", "nbinom", "zinb", "exp", "gamma", "lnorm", "gev", "ZInormal", "ZIgamma".

Author

Aimee Cody

Details

This function uses the forward-backward algorithm to compute local state probabilities from a fitted Hidden Markov Model (HMM). The most probable state at each observation is returned. Initial state probabilities are derived from the stationary distribution of the transition matrix Pi if not explicitly provided.

Supported observation distributions include normal, Poisson, Weibull, zero-inflated Poisson (ZIP), negative binomial, zero-inflated negative binomial (ZINB), exponential, gamma, log-normal, generalized extreme value (GEV), zero-inflated normal, and zero-inflated gamma.

See Also

findmleHMM, for estimating HMM parameters from data. generateHMM, for simulating HMM data. localdecodeHSMM, for local decoding of Hidden Semi-Markov Models. globaldecodeHMM, for global decoding in HMMs.

Examples

# Example with 3 states, normal observations
J <- 3

# HMM transition matrix
Pi <- matrix(c(0.8, 0.15, 0.05,
               0.1, 0.7, 0.2,
               0.2, 0.3, 0.5), nrow = 3, byrow = TRUE)

# Observation parameters (normal distribution)
obspar <- list(
  mean = c(-2, 0, 3),
  sd = c(0.5, 1, 1.5)
)

# Simulate HMM data
sim_data <- generateHMM(n = 200, J = J, obsdist = "norm", obspar = obspar, Pi = Pi)

# Fit HMM to simulated data
HMM_fit <- findmleHMM(J = J, x = sim_data$x, obsdist = "norm",
                       obspar = obspar, Pi = Pi, EM = FALSE)

# Compute local decoding
decoded_states <- localdecodeHMM(x = sim_data$x, HMM = HMM_fit, obsdist = "norm")

decoded_states