rho_IS_multivariate: rho Importance Sampling Step (multivariate)

In rchan26/hierarchicalFusion: Divide-and-Conquer Monte Carlo Fusion

rho Importance Sampling Step (multivariate)

Description

Performs the importance sampling step for rho where target is univariate

Usage

rho_IS_multivariate(
  particles_to_fuse,
  dim,
  N,
  m,
  time,
  inv_precondition_matrices,
  inverse_sum_inv_precondition_matrices,
  number_of_steps = 2,
  time_mesh = NA,
  resampling_method = "multi",
  seed = NULL,
  n_cores = parallel::detectCores(),
  cl = NULL
)

Arguments

particles_to_fuse	list of length m, where particles_to_fuse[[c]] contains the particles for the c-th sub-posterior (a list of particles to fuse can be initialised by initialise_particle_sets() function)
dim	dimension of the particles
N	number of particles to importance sample
m	number of sub-posteriors to combine
time	end time T for fusion algorithm
inv_precondition_matrices	list of length m of inverse preconditioning matrices
inverse_sum_inv_precondition_matrices	the inverse of the sum of the inverse precondition matrices (can be calculated by passing the inverse precondition matrices into inverse_sum_matrices())
number_of_steps	integer value for number of steps in the Fusion algorithm (default is 2 for Monte Carlo Fusion)
time_mesh	vector of times used in Fusion algorithm (default is NA). If set to NA, the returned particle has time_mesh given by c(0, time)
resampling_method	method to be used in resampling, default is multinomial resampling ('multi'). Other choices are stratified resampling ('strat'), systematic resampling ('system'), residual resampling ('resid')
seed	seed number - default is NULL, meaning there is no seed
n_cores	number of cores to use
cl	an object of class "cluster" for parallel computation in R. If none is passed, then one is created and used within this function

Value

A importance weighted particle set

Examples

samples_to_fuse <- lapply(1:2, function(i) mvrnormArma(100, c(0, 0), diag(2)))
particles_to_fuse <- initialise_particle_sets(samples_to_fuse = samples_to_fuse,
                                              multivariate = TRUE)
precondition_mats <- lapply(samples_to_fuse, cov)
inv_precondition_mats <- lapply(precondition_mats, solve)
inv_sum_inv_precondition_mats <- inverse_sum_matrices(inv_precondition_mats)
particles <- rho_IS_multivariate(particles_to_fuse = particles_to_fuse,
                                 N = 100,
                                 dim = 2,
                                 m = 2,
                                 time = 0.5,
                                 inv_precondition_matrices = inv_precondition_mats,
                                 inverse_sum_inv_precondition_matrices = inv_sum_inv_precondition_mats)

rchan26/hierarchicalFusion documentation built on Sept. 11, 2022, 10:30 p.m.