construct.emission.setup: Get an emission distribution design setup object

In cthmm: Latent Continuous Time Markov Chains

Description

The design matrix needs to all i,j entries in the emission matrix that are non-zero, except for the reference state for row i.

Usage

  construct.emission.setup(emission.states = NULL,
    ref.states = NULL, exact.states = NULL,
    fixed.dist = NULL, design.matrix = NULL,
    covariate.array = NULL, deriv.array = NULL,
    param.values = NULL, param.types = NULL,
    variable.predictors = NULL)

Arguments

emission.states	number of hidden states
ref.states	vector of index of states with non-zero initial probabilities
exact.states	index of reference state
fixed.dist	option emission matrix set at fixed value for all individuals.
design.matrix	array of dim(num beta parameters, num.emissions) that encodes the variables/constants used in linear predictors of log(emission_ij). Variables are named, and constants are entered as constants. The rows correspond to the transitions in the transition.codes entry.
covariate.array	(dim=c(num.betas,num.thetas, num.subjects)) with covariates values for all subjects
deriv.array
param.values	inital or fixed values of beta parameters
param.types	a vector of length(num.beta.parameters), with entries=0 if parameter is free, 1 if fixed at param.value.
variable.predictors	matrix with columns "i" and "j" that encode the variable predictors in the design.matrix

Value

emission.setup object (a list)

Author(s)

Jane Lange

cthmm documentation built on May 2, 2019, 5:46 p.m.