smcparameters: SMC Tuning Parameters

Description Objects from the Class Slots Methods Author(s) See Also Examples

Description

This class holds parameters for the Sequential Monte Carlo sampler.

Objects from the Class

Objects can be created by calls of the function "smcparameters".

Slots

nparticles:

Object of class "numeric": an integer representing the desired number of particles.

temperatures:

Object of class "numeric": a vector of temperatures, default being "seq(from = 0.01, to = 1, length.out = 100)".

nmoves:

Object of class "numeric": number of move steps to be performed after each resampling step, default being 1.

ESSthreshold:

Object of class "numeric": resampling occurs when the Effective Sample Size goes below "ESSthreshold" multiplied by the number of particles "nparticles".

movetype:

Object of class "character": type of Metropolis-Hastings move step to be performed; can be either set to "independent" or "randomwalk", default being "independent".

movescale:

Object of class "numeric": if movetype is set to "randomwalk", this parameter specifies the amount by which the estimate of the standard deviation of the target distribution is multiplied; the product being used to propose new points in the random-walk MH step. Default is 10%, ie a new point is proposed from a Normal distribution, centered on the latest point, with standard deviation equal to 10% of the standard deviation of the already-generated chain.

resamplingscheme:

Object of class "character": type of resampling to be used; either "multinomial", "residual" or "systematic", the default being "systematic".

Methods

show

signature(object = "smcparameters"): provides a little summary of a binning object when called (or when print is called).

Author(s)

Luke Bornn <bornn@stat.harvard.edu>, Pierre E. Jacob <pierre.jacob.work@gmail.com>

See Also

smc

Examples

1
2
3
4
5
showClass("smcparameters")
smcparam<- smcparameters(nparticles=5000, 
                        temperatures = seq(from = 0.0001, to = 1, length.out= 100),
                        nmoves = 5, ESSthreshold = 0.5, movetype = "randomwalk",
                        movescale = 0.1)

PAWL documentation built on May 2, 2019, 5:58 a.m.