weighted_logLik: Tune and evaluate weighted transformation ensembles

In deeptrafo: Fitting Deep Conditional Transformation Models

Tune and evaluate weighted transformation ensembles

Description

Tune and evaluate weighted transformation ensembles

Usage

weighted_logLik(
  object,
  weights = NULL,
  newdata = NULL,
  convert_fun = function(x, ...) mean(x, ...),
  batch_size = NULL,
  ...
)

Arguments

object	Object of class "dtEnsemble"
weights	Numeric; weight-vector of length n_ensemble, if NULL the weights are tuned on newdata
newdata	List or data.frame; new data to evaluate or tune the weights on
convert_fun	Function; applied to the log-likelihood values of all observations.
batch_size	Integer; optional, useful if data is too large.
...	Further arguments supplied to print.deeptrafo

Value

Returns list of ensemble members, average, and ensemble log-likelihood converted by convert_fun

deeptrafo documentation built on April 3, 2025, 10:38 p.m.