Static class with multiple, different evaluation functions. These functions can be used to either deterimine the loss of an estimator, or to determine its performance. Evaluation.get_evaluation_function

Evaluation.get_evaluation_function(family, useAsLoss = TRUE)

Arguments

family

the family of the data (binomial for binary, gaussian for cts).

useAsLoss

should we use the loss function or the performance function?

Value

function the log likelihoodloss function if useAsLoss = TRUE, and the log loss function otherwise.

Details

Returns the correct evaluation function given the family of the data. # @import Metrics