Function ecv.regression.baselearner.control sets up the base learners used in the ecv.regression call.
ecv.regression.baselearner.control(
baselearners = c("nnet", "rf", "svm", "gbm", "knn", "penreg")
, baselearner.configs = make.configs(baselearners, type = "regression")
, npart = 1, nfold = 5
)
ecv.regression.integrator.control(errfun=rmse.error, method=c("default"))Both functions return lists with same element names as function arguments.
Names of base learners used. Currently, regression options available are Neural Network ("nnet"), Random Forest ("rf"), Support Vector Machine ("svm"), Gradient Boosting Machine ("gbm"), and K-Nearest Neighbors ("knn"), Penalized Rergession ("penreg") and Bayesian Additive Regression Trees ("bart"). The last learner is not included by default, due to significantly longer training time needed by it ("bart") compared to other learners.
List of base learner configurations. Default is to call make.configs from package EnsembleBase.
Number of partitions to train each base learner configuration in a CV scheme.
Number of folds within each data partition.
Error function used to compare performance of base learner configurations. Default is to use rmse.error from package EnsembleBase.
Integrator method. Currently, only option is "default", which uses average error for each base learner configuration across repeated CV runs to chose the best configuration.
Mansour T.A. Sharabiani, Alireza S. Mahani
make.configs, rmse.error