logloss.md

logloss	R Documentation

Logarithmic Loss

Description

A generic S3 function to compute the logarithmic loss score for a classification model. This function dispatches to S3 methods inlogloss() and performs no input validation. If you supply NA values or vectors of unequal length (e.g. length(x) != length(y)), the underlying C++ code may trigger undefined behavior and crash your R session.

Defensive measures

Because logloss() operates on raw pointers, pointer-level faults (e.g. from NA or mismatched length) occur before any R-level error handling. Wrapping calls in try() or tryCatch() will not prevent R-session crashes.

To guard against this, wrap logloss() in a "safe" validator that checks for NA values and matching length, for example:

safe_logloss <- function(x, y, ...) {
  stopifnot(
    !anyNA(x), !anyNA(y),
    length(x) == length(y)
  )
  logloss(x, y, ...)
}

Apply the same pattern to any custom metric functions to ensure input sanity before calling the underlying C++ code.

Usage

## Generic S3 method
## for Logarithmic Loss
logloss(...)

## Generic S3 method
## for weighted Logarithmic Loss
weighted.logloss(...)

Arguments

...

Arguments passed on to logloss.integer,logloss.factor, weighted.logloss.integer,weighted.logloss.factor actual A vector length n, and k levels. Can be of integer or factor. response A n \times k <double>-matrix of predicted probabilities. The i-th row should sum to 1 (i.e., a valid probability distribution over the k classes). The first column corresponds to the first factor level in actual, the second column to the second factor level, and so on. normalize A <logical>-value (default: TRUE). If TRUE, the mean cross-entropy across all observations is returned; otherwise, the sum of cross-entropies is returned. w A <double> vector of sample weights.

Value

A <double>

References

MacKay, David JC. Information theory, inference and learning algorithms. Cambridge university press, 2003.

Kramer, Oliver, and Oliver Kramer. "Scikit-learn." Machine learning for evolution strategies (2016): 45-53.

Virtanen, Pauli, et al. "SciPy 1.0: f'undamental algorithms for scientific computing in Python." Nature methods 17.3 (2020): 261-272.

Examples

## Classes and
## seed
set.seed(1903)
classes <- c("Kebab", "Falafel")

## Generate actual
## and predicted response
## probabilities
actual_classes <- factor(
x = sample(x = classes, size = 1e3, replace = TRUE),
levels = c("Kebab", "Falafel")
)

response <- runif(n = 1e3)

## Logloss
SLmetrics::logloss(
   actual    = actual_classes, 
   response  = cbind(
 response,
 1 - response
   )
)

## Generate observed
## frequencies 
actual_frequency <- sample(10L:100L, size = 1e3, replace = TRUE)

## Poisson Logloss
SLmetrics::logloss(
   actual    = actual_frequency, 
   response  = response
)

</div>