Create setting for gradient boosting machine model using gbm_xgboost implementation

setGradientBoostingMachine(ntrees = c(100, 1000), nthread = 20,
  earlyStopRound = 25, maxDepth = c(4, 6, 17), minRows = 2,
  learnRate = c(0.005, 0.01, 0.1), seed = NULL)

Arguments

ntrees

The number of trees to build

nthread

The number of computer threads to (how many cores do you have?)

earlyStopRound

If the performance does not increase over earlyStopRound number of interactions then training stops (this prevents overfitting)

maxDepth

Maximum number of interactions - a large value will lead to slow model training

minRows

The minimum number of rows required at each end node of the tree

learnRate

The boosting learn rate

seed

An option to add a seed when training the final model

Examples

model.gbm <- setGradientBoostingMachine(ntrees=c(10,100), nthread=20, maxDepth=c(4,6), learnRate=c(0.1,0.3))