PredictionStabilityCriterion<T>

Models & Types · AiDotNet.ActiveLearning.StoppingCriteria

Stopping criterion based on prediction stability across iterations.

Implements: IStoppingCriterion<T>

For Beginners

This criterion stops learning when the model’s predictions on the unlabeled pool stop changing significantly. Stable predictions indicate the model has converged and won’t benefit much from more labeled data.

How It Works

How It Works:

Stability Measures:

When to Use:

Constructors

Constructor Summary
PredictionStabilityCriterion() Initializes a new PredictionStability criterion with default parameters.
PredictionStabilityCriterion(double stabilityThreshold, int patience, StabilityMeasure measure) Initializes a new PredictionStability criterion with specified parameters.

Properties

Property Summary
T CurrentStability { get; } Gets the current stability measurement.
int StableIterations { get; } Gets the number of consecutive stable iterations.
string Description { get; }
string Name { get; }

Methods

Method Summary
T GetProgress(ActiveLearningContext<T> context)
bool ShouldStop(ActiveLearningContext<T> context)
void Reset()