hmm_logistic_diffusion_full: Construct hidden Markov model obtained by discretizing...

In jeremyhengjm/UnbiasedGradients: Unbiased estimators for partially observed diffusions

Construct hidden Markov model obtained by discretizing Lamperti transformed logistic diffusion process

Description

Define a hidden Markov model obtained by discretizing Lamperti transformed logistic diffusion process.

Usage

hmm_logistic_diffusion_full(times)

Arguments

times

vector specifying observation times

Value

a list with objects such as: xdimension is the dimension of the latent process; ydimension is the dimension of the observation process; theta_dimension is the dimension of the parameter space; theta_names is a vector of characters to enumerate the parameters; theta_positivity is a vector logicals to index parameters with positivity constraints; construct_discretization outputs a list containing stepsize, nsteps, statelength and obstimes; construct_successive_discretization outputs lists containing stepsize, nsteps, statelength, obstimes for fine and coarse levels, and coarsetimes of length statelength_fine indexing time steps of coarse level; sigma is the diffusion coefficient of the process; rinit to sample from the initial distribution; rtransition to sample from the Markov transition; dtransition to evaluate the transition density; dmeasurement to evaluate the measurement density; functional is the smoothing function to compute gradient of log-likelihood.

jeremyhengjm/UnbiasedGradients documentation built on Nov. 19, 2023, 11:24 p.m.