| gibbs {RGeostats} | R Documentation |
Perform simulations using gibbs sampler
gibbs(db = NA, model = model.input(), seed = 232131, nbsimu = 1, rank.grf = 1, nboot = 10, niter = 100, flag.norm=FALSE, flag.propagation=FALSE, nostat=NA, percent=0.1, flag.ce=FALSE, flag.cstd=FALSE, verbose=FALSE, radix = "Gibbs", modify.target = db.locmod())
db |
The |
model |
The |
seed |
Seed used for the random number generator. When 0, the seed is not initialized. |
nbsimu |
Number of simulations |
rank.grf |
Rank of the gaussian random function to be simulated |
nboot |
Number of iterations performed for bootstrap |
niter |
Maximum number of iterations for calculating the conditional expectation |
flag.norm |
When TRUE, the model is normalized (as it should correspond to a normalized gaussian). |
flag.propagation |
When TRUE, the Gibbs sampler using the propagation algorithm is used. This method is incompatible with the presence of bounds. |
nostat |
List of non-stationary parameters. For details see
|
percent |
Amount of nugget effect added to the model if this initial model only contains Gaussian structures. If set to zero, the initial model is kept unchanged. This amount is defined as a percentage of the global variance. |
flag.ce |
When TRUE, the Conditional Expectation is derived from the simulations |
flag.cstd |
When TRUE, the Conditional Standard Deviation is derived from the simulations |
verbose |
Verbose flag |
radix |
Radix of the name given to the variables storing the results in the target Db. |
modify.target |
Decides whether or not the newly created variables will have their
locator defined or not. For more information, see |
Some parameters are passed by the 'keypair' mechanism (in the case of Gibbs Propagation algorithm):
gibbsPropaR: value of the coefficient 'r' used to generate the updated value. Default 0.
gibbsEps: the threshold (to consider that samples are not correlated) is calculated as the C(0) * gibbsEps. Default 0.
The data Db to which several variables have been added:
if 'flag.ce' is specified: the Conditional Expectation variable
if 'flag.cstd' is specified: the Conditional Standard Deviation variable
otherwise, one new variable per simulation