search for best hyperparameters for metaheuristic optimization

Share:

Description

search for best hyperparameters for metaheuristic optimization

Usage

1
2
3
metaheurhyper(gridclassobject = examplegrid, searchtype = "grid",
  nrohyperparams = 3, iterations = 10, cores = 1, nholdout = 2,
  trials = 3, model = "rpart")

Arguments

gridclassobject

(GridClass) created by setgrid function in preprocomb package, defaults to examplegrid

searchtype

(character) grid or random search

nrohyperparams

(integer) number of hyperparameters used in random search, between 1 and 5, default to 3

iterations

(integer) number of iterations done for a restart

cores

(integer) number of cores used in computation of classification accuracies holdout rounds

nholdout

(integer) number of holdout rounds in computing classification accuracies

trials

(integer) number of trials

model

(character) caret name of predictive model, defaults to "rpart"

Examples

1
## result <- metaheurhyper(cores=2, trials=2, iterations=30)