ShapleyValueFitDetector<T, TInput, TOutput>

Models & Types · AiDotNet.FitDetectors

A detector that evaluates model fit quality using Shapley values to determine feature importance.

Inheritance: FitDetectorBase<T, TInput, TOutput> → ShapleyValueFitDetector<T, TInput, TOutput>

For Beginners

This class helps you understand which features (input variables) in your model are most important for making predictions. It uses something called “Shapley values” from game theory to fairly distribute the “credit” for predictions among all your features.

Think of it like figuring out which players on a sports team contributed most to winning a game. Shapley values help determine if your model is:

  • Overfitting: Relying too much on just a few features (like a team depending only on one star player)
  • Underfitting: Not using features effectively (like a team not using any player’s strengths)
  • Good fit: Using features in a balanced way (like a well-coordinated team)

This detector will give you recommendations on how to improve your model based on this analysis.

Constructors

Constructor Summary
ShapleyValueFitDetector(ShapleyValueFitDetectorOptions options) Initializes a new instance of the ShapleyValueFitDetector class.

Methods

Method Summary
FitDetectorResult<T> DetectFit(ModelEvaluationData<T, TInput, TOutput> evaluationData) Analyzes the model’s fit using Shapley values and returns detailed results.