IScoringRule<T>

Interfaces · AiDotNet.Scoring

Defines a proper scoring rule for evaluating probabilistic predictions.

Known implementations: ScoringRuleBase<T>

For Beginners

When you make a probabilistic prediction (like “70% chance of rain”), scoring rules tell you how good your prediction was after you see what actually happened. A proper scoring rule rewards you most for predicting probabilities that match reality - you can’t game the system by being overconfident or underconfident.

How It Works

A scoring rule measures how well a predicted probability distribution matches the observed outcomes. Proper scoring rules are maximized (or minimized) when the predicted distribution equals the true distribution, incentivizing honest forecasts.

Properties

Property Summary
bool IsMinimized { get; } Gets whether this scoring rule should be minimized (true) or maximized (false).
string Name { get; } Gets the name of this scoring rule.

Methods

Method Summary
T MeanScore(IParametricDistribution<T>[] distributions, Vector<T> observations) Computes the mean score over multiple prediction-observation pairs.
T Score(IParametricDistribution<T> distribution, T observation) Computes the score for a single prediction-observation pair.
Vector<T> ScoreGradient(IParametricDistribution<T> distribution, T observation) Computes the gradient of the score with respect to distribution parameters.