predict.booami: Predict from booami objects

In booami: Component-Wise Gradient Boosting after Multiple Imputation

Predict from booami objects

Description

Predict responses (link or response scale) from fitted booami models.

Usage

## S3 method for class 'booami_cv'
predict(object, newdata, type = c("link", "response"), ...)

## S3 method for class 'booami_pooled'
predict(object, newdata, type = c("link", "response"), ...)

## S3 method for class 'booami_multi'
predict(object, newdata, type = c("link", "response"), ...)

Arguments

object	A fitted booami object. One of: "booami_cv" — cross-validated model object. "booami_pooled" — pooled fit from impu_boost(..., pool = TRUE). "booami_multi" — unpooled fit from impu_boost(..., pool = FALSE).
newdata	A data.frame or matrix of predictors (same columns/order as training).
type	Either "link" for the linear predictor, or "response" for mean/probability (Gaussian/logistic respectively).
...	Passed to booami_predict. For "booami_multi", you may use aggregate = "mean"|"median"|NULL and/or which_m = <index> to control how predictions are aggregated across imputations.

Value

A numeric vector of predictions.

See Also

booami_predict

booami documentation built on Feb. 19, 2026, 5:07 p.m.