predict.GradientBoost: 'predict.GradientBoost'

In Displayr/flipMultivariates: Multivariate models

predict.GradientBoost

Description

Predicts values for numeric outcomes and group membership for categories based on newdata and a fitted GradientBoost object. A value (which may be NA) is returned for every instance including those with missing data. If newdata is not specified the fitted data before filtering are used. NA is returned for cases with unfitted factor levels or any missing data.

Usage

## S3 method for class 'GradientBoost'
predict(object, newdata = NULL, keep.soft.probs = FALSE, ...)

Arguments

object	A GradientBoost object.
newdata	Optionally, a data frame including the variables used to fit the model. If omitted, the data supplied to GradientBoost() is used before any filtering.
keep.soft.probs	A logical to control whether output returns soft probabilities (i.e. probabilities for each categorical level) are kept or output is coerced to a factor via the largest probability.
...	Additional arguments to pass to predict.xgb.Booster.

Displayr/flipMultivariates documentation built on Feb. 26, 2024, 12:39 a.m.