norm0_working_params: Transform normal HMM parameters from natural to working

In simonecollier/lizardHMM: Fitting Lizard Movement Data with HMMs

Description

The function transforms the natural normal HMM parameters that have additional constraints into working parameters that incorporate the constraints.

Usage

norm0_working_params(
  num_states,
  num_variables,
  num_subjects,
  mu,
  sigma,
  zweight,
  beta,
  delta
)

Arguments

num_states	The number of states in the desired HMM.
num_variables	The number of variables in the data.
num_subjects	The number of subjects/trials that generated the data.
mu	A list of matrices containing the means of the state dependent normal distribution. Each matrix corresponds to a different variable, each row corresponds to a different subject and each column corresponds to a different state.
sigma	A list of matrices containing the standard deviations of the state dependent normal distribution. Each matrix corresponds to a different variable, each row corresponds to a different subject and each column corresponds to a different state.
beta	A matrix of regression coefficients for the effect of the covariates on the transition probability matrix gamma.
delta	A list of the initial state distributions for each subject.

Value

A single vector containing working parameters.

Examples

# define values of parameters
num_states    <- 2
num_variables <- 2
num_subjects  <- 2
mu            <- list(matrix(c(1, 5, 2, 4), 2, 2, byrow = TRUE),
                      matrix(c(1, 5, 2, 4), 2, 2, byrow = TRUE))
sigma         <- list(matrix(c(1, 2, 1, 1.5), 2, 2, byrow = TRUE),
                      matrix(c(1, 2, 1, 1.5), 2, 2, byrow = TRUE))
beta          <- matrix(c(-2, 0, 0), nrow = 1, ncol = 3)
delta         <- list(c(1/2, 1/2), c(1/2, 1/2))

#transform to working parametes
norm_working_params(num_states, num_variables, num_subjects,
                    mu, sigma, beta, delta)

simonecollier/lizardHMM documentation built on Dec. 23, 2021, 2:24 a.m.