df_logloss: Logarithmic Loss (Logloss) (computation function, any size)

In Laurae2/Laurae: Advanced High Performance Data Science Toolbox for R

Description

This function computes the Logarithmic Loss (Logloss) provided preds and labels, while handling multiclass problems.

Usage

df_logloss(preds, labels, eps = 1e-15)

Arguments

preds	The predictions.
labels	The labels.
eps	The threshold for predictions so you do not have infinite loss issued from an observation. Defaults to 1e-15.

Value

The Logarithmic Loss.

Examples

library(data.table)

# Binary classification problem
my_preds <- dnorm(rnorm(n = 9*6*10))
my_labels <- as.numeric(runif(n = 9*6*10) >= 0.5)
df_logloss(my_preds, my_labels)

# Multiclass classification problem
my_preds <- data.table(matrix(dnorm(rnorm(n = 9*6*10)), nrow = 9*10))
my_labels <- rep(c(1, 2, 3, 4, 5, 1, 0, 2, 3), 10)
df_logloss(my_preds, my_labels)

Laurae2/Laurae documentation built on May 8, 2019, 7:59 p.m.