fastfun-class: Class for fast to compute objective.

In DiceOptim: Kriging-Based Optimization for Computer Experiments

Description

Class for fast to compute objective.

Usage

## S4 method for signature 'fastfun'
predict(object, newdata, ...)

## S4 method for signature 'fastfun'
update(object, newX, newy, ...)

## S4 method for signature 'fastfun'
simulate(object, nsim, seed, newdata, cond, nugget.sim, checkNames, ...)

Arguments

object	fastfun object
newdata	an optional vector, matrix or data frame containing the points where to perform predictions. Default is NULL: simulation is performed at design points specified in object.
...	further arguments (not used)
newX	Matrix of the new location for the design
newy	Matrix of the responses at newX
nsim	an optional number specifying the number of response vectors to simulate. Default is 1.
seed	usual seed argument of method simulate. Not used.
cond	an optional boolean indicating the type of simulations. Not used.
nugget.sim	an optional number corresponding to a numerical nugget effect. Not used.
checkNames	an optional boolean. Not used.

Methods (by generic)

• predict: Predict(by evaluating fun) the result at a new observation.

• update: Update the X and y slots with a new design and observation.

• simulate: Simulate responses values (for compatibility with methods using DiceKriging simulate)

Slots

d

spatial dimension,

n

observations number,

X

the design of experiments, size n x d,

y

the observations, size n x 1,

fun

the evaluator function.

Objects from the Class

To create a fastfun object, use fastfun. See also this function for more details and examples.

DiceOptim documentation built on Feb. 2, 2021, 1:06 a.m.