Get_Rdiag: Checking convergence: diagnostics

In sppmix: Modeling Spatial Poisson and Related Point Processes

Description

This function reports the Gelman-Rubin convergence diagnostic R (also known as the potential scale reduction), by producing k DAMCMC fits and computing the within-chain and between-chain variances. Values approximately equal to 1 indicate convergence, otherwise we need to run the chain for a longer number of iterations to get convergence.

For DAMCMC examples see

http://faculty.missouri.edu/~micheasa/sppmix/sppmix_all_examples.html#Get_Rdiag

Usage

Get_Rdiag(pp, m, truncate = FALSE, L = 20000, numofchains = 2,
  permute = 0)

Arguments

pp	Point pattern object of class ppp.
m	The number of components to fit.
truncate	Logical variable indicating whether or not we normalize the densities of the mixture components to have all their mass within the window defined in the point pattern pp.
L	Number of iterations to use for each chain created; default is 20000. Note that half of them will be dropped so use a large number.
numofchains	Number of chains to create; default is 2.
permute	Request to generate chains that are unpermuted (permute=0), identifiability constraint (IC) permuted (permute=1), or minimum squared error loss (SEL) permuted (permute=2).

Author(s)

Sakis Micheas

See Also

est_mix_damcmc, rmixsurf, rsppmix

Examples

truemix_surf <- rmixsurf(m = 3, lambda=100, xlim = c(-3,3), ylim = c(-3,3))
plot(truemix_surf)
genPPP=rsppmix(intsurf = truemix_surf, truncate = FALSE)
Get_Rdiag(pp = genPPP, m = 3)

sppmix documentation built on Jan. 13, 2021, 10:04 p.m.