cma_jTuner: Perform CMA-ES tuning (Java version).

In TDMR: Tuned Data Mining in R

Description

Perform a parameter tuning by CMA-ES, using the *Java* implementation by Niko Hansen through the interface package rCMA-package.

Usage

cma_jTuner(confFile, tdm, envT, dataObj)

Arguments

confFile	task configuration for tuning algorithm
tdm	the list with TDMR settings. If tdm$CMA.propertyFile is not NULL, then rCMA-package will read this CMA property file. If NULL, the default file CMAEvolutionStrategy.properties from find.package("rCMA") is taken. If tdm$CMA.populationSize is not NULL, the CMA population size will be set accordingly. If NULL, the default population size 4 + 3*log(N) is taken with N = number of parameters in .roi file.
envT	the environment variable
dataObj	the TDMdata object containing the data set (train/vali part and test part)

Value

the result of CMA-ES tuning, i.e. the list envT$spotConfig, extended by list cma with elements

cma$sres	a string vector with all console output from the optimization run cmaOptimDP of package rCMA-package
cma$count	the number of calls to tdmStartOther

TDMR documentation built on March 3, 2020, 1:06 a.m.