VariationRatios<T>
Models & Types · AiDotNet.ActiveLearning
Implements variation ratios for active learning sample selection.
Implements: IActiveLearningStrategy<T>
For Beginners
Variation ratios measure the proportion of predicted labels that are NOT in the modal (most likely) class. High variation ratio indicates high uncertainty because the model is not strongly predicting any single class.
How It Works
Formula: VR(x) = 1 - max(P(y|x))
Interpretation:
Advantages:
Reference: Freeman, L.C. (1965). “Elementary Applied Statistics: For Students in Behavioral Science.”
Constructors
| Constructor | Summary |
|---|---|
VariationRatios() |
Initializes a new instance of the VariationRatios class. |
Properties
| Property | Summary |
|---|---|
bool UseBatchDiversity { get; set; } |
|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
Dictionary<string, T> GetSelectionStatistics() |
|
Vector<T> ComputeInformativenessScores(IFullModel<T, Tensor<T>, Tensor<T>> model, Tensor<T> unlabeledPool) |
|
int[] SelectSamples(IFullModel<T, Tensor<T>, Tensor<T>> model, Tensor<T> unlabeledPool, int batchSize) |