gelman_diag: Convergence check by Gelman's potential scale reduction.
In ddepn: Dynamic Deterministic Effects Propagation Networks: Infer signalling networks for timecourse RPPA data.

Description Usage Arguments Details Value Note References See Also Examples

Compute Gelman's potential scale reduction to estimate convergence in inhibMCMC.

1	gelman_diag(ret)

ret

List. Object returned by ddepn, when inhibMCMC was used as inference method.

The function is a wrapper calling the gelman.diag function from package coda.

The result of the gelman.diag function.

Note that multiple chains have to be performed to use this diagnostic.

Gelman, A., Carlin, JB., Stern, HS., Rubin, DB.: Bayesian Data Analysis, 2nd edition, Chapman & Hall/CRC, chapter 11.6, pp294-295.

coda package.

## Not run: 
library(ddepn)
set.seed(12345)
n <- 6
signet <- signalnetwork(n=n, nstim=2, cstim=0, prop.inh=0.2)
net <- signet$phi
stimuli <- signet$stimuli
weights <- signet$weights

## sample data
dataset <- makedata(net, stimuli, mu.bg=1200, sd.bg=400, mu.signal.a=2000, sd.signal.a=1000)
data <- dataset$datx

# use the original network as prior probability matrix
B <- net
B[B==2] <- -1
# construct a prior matrix with uniform probabilities for each edge
if(require(multicore)) {
  ret <- ddepn(data, phiorig=net, inference="mcmc",
    maxiterations=3000, burnin=1000,
    usebics=FALSE, lambda=0.01, B=B,
    multicores=TRUE, cores=4, priortype="laplaceinhib")
}

## now produce the convergence diagnostic
gelman_diag(ret)



## End(Not run)