PoolBasedSplitter<T>

Models & Types ยท AiDotNet.Preprocessing.DataPreparation.Splitting.ActiveLearning

Pool-based active learning splitter that maintains labeled and unlabeled pools.

Inheritance: DataSplitterBase<T> โ†’ PoolBasedSplitter<T>

For Beginners

In pool-based active learning, we start with a small labeled dataset and a large pool of unlabeled data. The model queries samples from the unlabeled pool that it is most uncertain about, and an oracle (human expert) provides labels.

How It Works

How It Works:

  1. Start with a small initial labeled set (seed)
  2. The rest becomes the unlabeled pool
  3. Model queries uncertain samples from the pool
  4. Oracle labels them, moving them to the labeled set

When to Use:

  • Limited labeling budget
  • Expensive manual annotation
  • Interactive machine learning
  • Iterative model improvement

Constructors

Constructor Summary
PoolBasedSplitter(double initialLabeledRatio, int minLabeledSamples, SelectionStrategy<T> strategy, bool shuffle, int randomSeed) Creates a new pool-based active learning splitter.

Properties

Property Summary
bool RequiresLabels { get; }
string Description { get; }

Methods

Method Summary
DataSplitResult<T> Split(Matrix<T> X, Vector<T> y)