ParticleMatrix: Class "ParticleMatrix"

In SMCS4: Sequential Monte Carlo samplers in S4

Description

Basic class to implement a Sequential Monte Carlo sampler.

Slots

Code: N lW_update logLik logWeights mcmc_move p_margin p_move particles resampleC unifWeights Inherited: particles logWeights unifWeights p_move mcmc_move lW_update logLik resampleC N Docs: Particles codep_margin: lW_update logWeights mcmc_move p_move resampleC unifWeights

particles:

Matrix with particle values

p_margin:

Integer indicating which margin in the matrix represents particle ids

logWeights:

Vector containing the log (unnormalised) particle weights.

unifWeights

Logical indicating whether the logWeights are uniform.

p_move:

Function to move the particles to a new position.

mcmc_move:

Function to perform a Monte Carlo Markov Chain move.

lW_update:

Function to do update the logWeights.

resampleC:

Numeric value (between 0 and 1) indicating when to perform resampling.

Methods

particles

signature(object = "ParticleMatrix"): return particles

particles<-

signature(object = "ParticleMatrix"): set particles

ParticleMove

signature(object = "ParticleMatrix"): move particles

#

doParticleMove

signature(object = "ParticleMatrix"): move particles

SmcIterate

signature(object = "ParticleMatrix"): do a full SMC iteration

#

doSmcIterate

signature(object = "ParticleMatrix"): move particles

UpdateWeights

signature(object = "ParticleMatrix"): calculate updated weights

#

doUpdateWeights

signature(object = "ParticleMatrix"): move particles

ESS

signature(object = "ParticleMatrix"): effective sample size

logWeights

signature(object = "ParticleMatrix"): return log weights

logWeights<-

signature(object = "ParticleMatrix"): set log weights

getWeights

signature(object = "ParticleMatrix"): return the unnormalized weights

getNormWeights

signature(object = "ParticleMatrix"): return the self-normalized weights

mean

signature(object = "ParticleMatrix"): return the estimated mean of each particle dimension

var

signature(object = "ParticleMatrix"): return the estimated variance of each particle dimension

cov

signature(object = "ParticleMatrix"): return the estimated covariance matrix of the particle dimensions

Author(s)

Maarten Speekenbrink

SMCS4 documentation built on May 2, 2019, 6:40 p.m.