HoldoutValidationFitDetector<T, TInput, TOutput>

Models & Types · AiDotNet.FitDetectors

A detector that evaluates model fit quality using holdout validation techniques.

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

For Beginners

This class helps you determine if your machine learning model is performing well by comparing how it performs on different subsets of your data:

  • Training data: The data used to build the model
  • Validation data: A separate set of data used to tune the model
  • Test data: A final set of data used to evaluate the model’s performance

By comparing performance across these sets, the detector can identify common problems like:

  • Overfitting: When your model performs very well on training data but poorly on new data
  • Underfitting: When your model performs poorly on all data sets
  • High variance: When your model’s performance varies significantly between different data sets

Constructors

Constructor Summary
HoldoutValidationFitDetector(HoldoutValidationFitDetectorOptions options) Initializes a new instance of the HoldoutValidationFitDetector<T> class.

Methods

Method Summary
FitDetectorResult<T> DetectFit(ModelEvaluationData<T, TInput, TOutput> evaluationData) Analyzes model performance data to detect the quality of fit.