Bergm (version 1.0)

bgof: Bayesian goodness-of-fit diagnostics

Description

Calculates summaries for degree, minimum geodesic distances, and edge-wise shared partner distributions to diagnose the Bayesian goodness-of-fit of exponential family random graph models.

Usage

bgof(out, directed = FALSE, lag = 100, n.sim = NULL, burn.in = 10000, n.deg = NULL, n.dist = NULL, n.esp = NULL, n.ideg = NULL, n.odeg = NULL, save = FALSE)

Arguments

out

bergm output

directed

logical; TRUE if the observed graph is directed

lag

count; interval between each sampled posterior parameter estimate

n.sim

count; number of sampled parameter values

burn.in

count; number of iterations used for each graph simulation

n.deg

count; used to plot only the first n.deg-1 degree distributions

n.dist

count; used to plot only the first n.dist-1 geodesic distances distributions

n.esp

count; used to plot only the first n.esp-1 edge-wise shared partner distributions

n.ideg

count; used to plot only the first n.ideg-1 in-degree distributions

n.odeg

count; used to plot only the first n.odeg-1 out-degree distributions

save

logical; if TRUE a file called "bgof.out" is saved in the working directory

Examples

data(florentine)

flo1 <- bergm(flobusiness~edges+kstar(2),main.iter=15000,
sdprop=c(1,0.1),sdprior=20,theta=c(-2,0.1))

## Bayesian goodness-of-fit test 

bgof(flo1,lag=150,n.sim=100)

## Let's improve the plots

bgof(flo1,lag=150,n.sim=100,n.deg=10,n.dist=9,n.esp=6)