predict.deeppam: Predicted values based on deeppam object.

In pkopper/deeppam: Machine Learning for Time-to-Event Data with Semi-Structured Deep Piecewise Additive Models

Description

Predicted values based on deeppam object.

Usage

## S3 method for class 'deeppam'
predict(
  object,
  x = NULL,
  output = c("tensor", "matrix", "numeric"),
  internal = TRUE,
  ...
)

Arguments

object	keras model of object class "deeppam".
x	non-optimal data for predicitons. Must be the same format as for fitting.
output	a character indicicating how the output should be shaped. "tensor" is the standard output, having dimensions nxtx1, "matrix" is reduced to nXt, "numeric2 is reduced to n*t.
internal	a logical indicating if the predicition is internal (used for e.g. computing validation loss) or external (e.g. for evaluation). This argument has effect on wheter the offset is considered when predicting. Thus, the default is TRUE so that during fitting this can be accessed.
...	further arguments passed to or from other methods.

Author(s)

Philipp Kopper

pkopper/deeppam documentation built on Jan. 19, 2021, 12:39 a.m.