cubfits (version 0.1-2)

Cedric Convergence Utilities: Cedric Convergence Utilities

Description

This utility function provides convergence related functions by Cedric.

Usage

cubmultichain(cubmethod, reset.qr, seeds=NULL, teston=c("phi", "sphi"), swap=0, swapAt=0.05, monitor=NULL,  min=0, max=160000, nchains=2, conv.thin=10, eps=0.1, ncores=2, ...)  cubsinglechain(cubmethod, frac1=0.1, frac2=0.5, reset.qr, seed=NULL, teston=c("phi", "sphi"), monitor=NULL,  min=0, max=160000, conv.thin=10, eps=1, ...)

Arguments

cubmethod

String to choose method. Options are "cubfits", "cubappr", "cubpred"

reset.qr

recalculate QR decomposition matrix of covariance matrix until reset.qr samples are reached

swap

proportion of b matrix parameters to be swaped between convergence checks

swapAt

difference (L1-norm) between two consequtive convergence test leading to a swap in the b matrix

seeds

Vector of seed for random number generation

seed

Seed for random number generation

teston

Select data to test convergence on

monitor

A function to monitor the progress of the MCMC. The fucntions expects the result object and for cubmultichain an index i. (cubmultichain call: monitor(x,i), cubsinglechain call: monitor(x))

min

Minimum samples to be obtained. eps is ignored until number of samples reaches min

max

Maximum samples to be obtained. eps is ignored after max samples is obtained

eps

Convergence criterium

conv.thin

thinning of samples before performing convergence test

nchains

number of chains to run in parallel

ncores

number of cores to use for parallel execution of chains

frac1

fraction of samples at the beginning of set for Geweke test

frac2

fraction of samples at the end of set for Geweke test

...

named arguments for functions "cubfits", "cubappr" or "cubpred"

Value

under development

Details

under development

References

https://github.com/clandere/cubfits/

See Also

cubfits, cubappr, cubpred

Examples

## Not run: 
# suppressMessages(library(cubfits, quietly = TRUE))
# 
# ## End(Not run)