MLmetrics (version 1.0.0)

MultiLogLoss: Multi Class Log Loss

Description

Compute the multi class log loss.

Usage

MultiLogLoss(y_true, y_pred)

Arguments

y_true

Ground truth (correct) labels vector or a matrix of correct labels indicating by 0-1, same format as probabilities matrix

y_pred

Predicted probabilities matrix, as returned by a classifier

Value

• Multi Class Log Loss

Examples

data(iris)
svm.model <- e1071::svm(Species~., data=iris, probability=TRUE)
pred <- predict(svm.model, iris, probability=TRUE)
MultiLogLoss(y_true=iris$Species, y_pred=attr(pred, "probabilities"))