update {DiceKriging}

Update of a kriging model

Description

Update a km object when one or many new observations are added. Many, but not all, fields of the km object need to be recalculated when new observations are added. It is also possible to modify the k last (existing) observations.

Usage

## S4 method for signature 'km'
update(object, newX, newy, newX.alreadyExist = FALSE,
          cov.reestim = TRUE, trend.reestim = TRUE, nugget.reestim = FALSE, 
          newnoise.var = NULL, kmcontrol = NULL, newF = NULL,...)

Arguments

object

Kriging model of km class.

newX

Matrix with object@d columns and r rows corresponding to the r locations of the observations to be updated. These locations can be new locations or existing ones.

newy

Matrix with one column and r rows corresponding to the r responses at the r locations newX.

newX.alreadyExist

Boolean: indicate whether the locations newX are all news or not.

cov.reestim

Should the covariance parameters of the km object be re-estimated?

trend.reestim

Should the trend parameters be re-estimated?

nugget.reestim

Should the nugget effect be re-estimated?

newnoise.var

Vector containing the noise variance at each new observations.

kmcontrol

Optional list representing the control variables for the re-estimation of the kriging model once new points are sampled. The items are the same as in km

newF

Optional matrix containing the value of the trend at the new locations. Setting this argument avoids a call to an expensive function.

...

Further arguments

Value

Updated km object

Author(s)

Clement Chevalier (IMSV, Switzerland, and IRSN, France)

References

Bect J., Ginsbourger D., Li L., Picheny V., Vazquez E. (2010), Sequential design of computer experiments for the estimation of a probability of failure, Statistics and Computing, pp.1-21, 2011, https://arxiv.org/abs/1009.5177

Chevalier C., Bect J., Ginsbourger D., Vazquez E., Picheny V., Richet Y. (2011), Fast parallel kriging-based stepwise uncertainty reduction with application to the identification of an excursion set, https://hal.archives-ouvertes.fr/hal-00641108/

See Also

km

Examples

set.seed(8)
N <- 9 # number of observations
testfun <- branin

# a 9 points initial design 
design <- expand.grid(x1=seq(0,1,length=3), x2=seq(0,1,length=3))
response <- testfun(design)

# km object with matern3_2 covariance
# params estimated by ML from the observations
model <- km(formula = ~., design = design, 
	response = response, covtype = "matern3_2")
model@covariance

newX <- matrix(c(0.4,0.5), ncol = 2) #the point that we are going to add in the km object
newy <- testfun(newX)
newmodel <- update(object = model, newX = newX, newy = newy, cov.reestim = TRUE)
newmodel@covariance