recalculateParetoSet: Recalculate Pareto-optimal solutions

In TunePareto: Multi-Objective Parameter Tuning for Classifiers

Recalculate Pareto-optimal solutions

Description

Recalculates the Pareto-optimal solutions in a TuneParetoResult according to the specified objectives only, and returns another TuneParetoResult object reduced to these objectives. This avoids time-consuming recalculations of objective values if only a subset of objectives should be considered for a previously evaluated set of parameter combinations.

Usage

recalculateParetoSet(tuneParetoResult, 
                     objectives)

Arguments

tuneParetoResult	The TuneParetoResult object containing the parameter configurations to be examined
objectives	A vector of objective function indices. The Pareto set is recalculated according to these objectives, i.e. omitting other objectives. If this argument is not supplied, all objectives are used, which usually returns a copy of the input.

Value

Returns a reduced TuneParetoResult object. For more details on the object structure, refer to tunePareto.

See Also

tunePareto, mergeTuneParetoResults

Examples

# optimize the 'cost' parameter of an SVM on
# the 'iris' data set
res <- tunePareto(classifier = tunePareto.svm(),
                  data = iris[, -ncol(iris)], 
                  labels = iris[, ncol(iris)],
                  cost=seq(0.01,0.1,0.01),
                  objectiveFunctions=list(cvWeightedError(10, 10),
                                          cvSensitivity(10, 10, caseClass="setosa"),
                                          cvSpecificity(10, 10, caseClass="setosa")))
print(res)

# select only specificity and sensitivity
print(recalculateParetoSet(res, 2:3))

TunePareto documentation built on Oct. 2, 2023, 5:06 p.m.