fit.gbdt: Fit a Gradient Boosted Decision Tree model

In jiho/soundclass: Signal classifier for sounds

Description

Fit a Gradient Boosted Decision Tree model

Usage

fit.gbdt(x, n.trees = 5000, shrinkage = 0.001, interaction.depth = 3,
  cv.folds = 5, ...)

Arguments

x	a data.frame with signal characteristics in the first columns and signal type (classification) in the last column
n.trees	see ?gbm
shrinkage	see ?gbm
interaction.depth	see ?gbm
cv.folds	see ?gbm
...	passed to gbm

See Also

gbm, summary.gbm, predict.gbm in package gbm

Examples

data(sirena)
head(sirena)
m <- fit.gbdt(x=sirena, n.trees=3, shrinkage=0.01, interaction.depth=1,
              n.cores=1)
# NB: The arguments values are caricatural here for the example to run
#     quickly enough and because the data is easy.
# NB: n.cores = 1 is necessary for examples to run on all machines. Feel
#     free to remove it and use more cores on your machine.
print(m)
summary(m)
pred <- predict(m, newdata=sirena)
head(pred)
(cm <- confusion_matrix(true=sirena$type, pred=pred$type))
confusion_stats(cm)

jiho/soundclass documentation built on June 5, 2019, 10:10 p.m.