fit2function: Model Fit to Function

In martinzaefferer/COBBS: Continuous Optimization Benchmarks By Simulation

Model Fit to Function

Description

This function produces an objective function with y=f(x) from a provided model fit. Important note: this function expects predict(object,newdata) to return a list. The object$target parameter is a string that determins which list item is returned by the created function. If not set (NULL), object$target is set to "y".

Usage

fit2function(object, infillCriterion = NULL)

Arguments

object	fit created by a modeling function, e.g., gaussianProcessRegression
infillCriterion	optional parameter, a function that accepts prediction results and a model object. The function should use these to alter the prediction result in a user desired way. For example turning the prediction results of a kriging model (mean and sd) into the expected improvement criterion

Value

a function in the style of y=f(x), which uses the fitted object to predict y for sample x.

martinzaefferer/COBBS documentation built on July 19, 2023, 4:12 a.m.