TCAVResult<T>

Models & Types · AiDotNet.Interpretability.Explainers

Represents the result of a TCAV analysis for a single concept.

For Beginners

This contains everything you need to know about how a concept influences predictions for a specific class.

Constructors

Constructor Summary
TCAVResult(string conceptName, int targetClass, double[] tcavScores, double meanScore, double standardDeviation, double pValue, bool isSignificant, double cavAccuracy, List<ConceptActivationVector<T>> cavs) Initializes a new TCAV result.

Properties

Property Summary
IReadOnlyList<ConceptActivationVector<T>> CAVs { get; } Gets all trained CAVs.
bool IsSignificant { get; } Gets whether the result is statistically significant.
double CAVAccuracy { get; } Gets the mean CAV classifier accuracy.
double MeanScore { get; } Gets the mean TCAV score across all runs.
double PValue { get; } Gets the p-value from statistical significance testing.
double StandardDeviation { get; } Gets the standard deviation of TCAV scores.
double[] TCAVScores { get; } Gets all TCAV scores from different CAV runs.
int TargetClass { get; } Gets the target class being analyzed.
string ConceptName { get; } Gets the name of the concept being tested.

Methods

Method Summary
string ToString() Returns a human-readable summary.