gibbs.test {RGeostats}R Documentation

Testing Gibbs convergence

Description

Testing Gibbs convergence

Usage

gibbs.test(db, model = model.input(), seed = 232131, niter = 10, radius=1,
           nbyter = 1, flag.edge = TRUE, flag.reset = TRUE, radix ="Gibbs",
           modify.target = db.locmod())

Arguments

db

A db-class object (corresponding to a Grid) where the Gibbs algorithm is programmed to test its convergence.

model

The model-class object containing the Model to be simulated using the Gibbs Sampler algorithm.

seed

Seed used for the generation of the random numbers.

niter

Number of iterations for the Gibbs Sampler. One iteration is a complete review of the simulation over all the nodes of the grid.

radius

This argument provides a simplified version of the neighborhood.

nbyter

A printout is generated when the number of iteration is equal to a multiple of the 'nbyter' argument.

flag.edge

When set to 1, the grid is transformed into a periodic data set.

For example, if the grid nodes along one direction are ranked from 1 to N. The data set is expanded by replacing node N+1 by node 1, node N+2 by 2, ...etc. Similarly node N replaces node 0, node N-1 replaces node -1, ...etc.

flag.reset

When set to 1, the simulation is performed from scratch. Otherwise, the simulation resumes from the values currently stored in the variable (with locator Z) of the db-class.

radix

Radix of the name given to the variable storing the result of the Gibbs simulation.

modify.target

Decides whether or not the newly created variables will have their locator defined or not. For more information, see db.locmod.

Details

This function is a test version and is not supposed to be called by standard used.

Value

The db-class where the simulated Gibbs variable has been added.


[Package RGeostats version 11.0.6 Index]