predict.coxkl: Predict Linear Predictors from a 'coxkl' Object

In survkl: Estimate Survival Data with Data Integration

Predict Linear Predictors from a coxkl Object

Description

Computes linear predictors for new data based on a fitted coxkl model. If eta is supplied, predictions are returned for those eta values; otherwise predictions are returned for all fitted etas. Linear interpolation is applied if an intermediate eta value is requested.

Usage

## S3 method for class 'coxkl'
predict(object, newz, eta = NULL, ...)

Arguments

object	A fitted model object of class "coxkl".
newz	A numeric matrix or data frame of new covariates (must match the dimension of the training design matrix used to fit the model).
eta	Optional numeric vector of eta value(s) for which to predict. If NULL, predictions for all fitted eta values are returned.
...	Additional arguments.

Details

The linear predictors are computed as as.matrix(newz) %*% beta.

Value

A numeric matrix of linear predictors with one column per eta (sorted ascending).

See Also

coef.coxkl

survkl documentation built on April 22, 2026, 1:08 a.m.