Learn R Programming

codadiags (version 1.0)

Markov chain Monte Carlo burn-in based on "bridge" statistics

Description

Markov chain Monte Carlo burn-in based on "bridge" statistics, in the way of coda::heidel.diag, but including non asymptotic tabulated statistics.

Copy Link

Version

Install

install.packages('codadiags')

Monthly Downloads

18

Version

1.0

License

GPL-3

Maintainer

Yann Richet

Last Published

November 18th, 2013

Functions in codadiags (1.0)

AR1

Generate auto-regressive order 1 sequence
null.param.cdf

Build the null CDF (cumulative density function) for a given statistic, for arbitrary length and autocorrelation sequence.
bay.cdf

Bay cumulative density function, corresponding to -B(t+)/B(t-), where B(t+) (resp. B(t-)) is the maximum (resp.minimum) of B(t)/(t*(1-t)).
transient.test

Perform a stationary test to check for an initial burn-in in a sequence
bridgestat.diag

Iterative truncation procedure based on a bridge statistic.
maxinv.bay.cdf

CDF of max(x,1/x) (=cdf(x)-cdf(1)+cdf(1)-cdf(1/x)) where x is 'Bay' distributed
null.lim.cdf

Asymptotic CDF for a given statistic
loglikbridge

Compute the so called "Log-likelihood bridge" process.
studentbridge

Compute the so called "Student bridge" process.
ad.cdf

Anderson-Darling cumulative density function, copy from ADGofTest package.
brownianbridge

Compute the so called (abusively) "Brownian bridge" process.
add.transient

Add a transient to a given mcmc sequence
ks.cdf

Kolmogorov-Smirnov cumulative density function, copy from stats::ks.test.
codadiags-package

Markov chain Monte Carlo burn-in based on "bridge" statistics.
cvm.cdf

Cramer von Mises cumulative density function, import from coda package.
autocorr1

Basic auto-correlation estimation of a given sequence