QueryByCommitteeSplitter<T>
Models & Types · AiDotNet.Preprocessing.DataPreparation.Splitting.ActiveLearning
Query-by-Committee (QBC) splitter for ensemble-based active learning.
Inheritance: DataSplitterBase<T> → QueryByCommitteeSplitter<T>
For Beginners
Query-by-Committee maintains multiple models (a “committee”) that are trained on different subsets of the labeled data. Samples where the committee disagrees most are the most informative and should be labeled next.
How It Works
How It Works:
- Create multiple subsets of the initial labeled data (one per committee member)
- Each subset can be used to train a different model
- Unlabeled samples with highest disagreement are prioritized for labeling
When to Use:
- When you can train multiple models
- Ensemble-based approaches
- Maximizing information gain from labeling
- Reducing labeling costs through disagreement sampling
Constructors
| Constructor | Summary |
|---|---|
QueryByCommitteeSplitter(int committeeSize, double subsampleRatio, double initialLabeledRatio, bool shuffle, int randomSeed) |
Creates a new Query-by-Committee splitter. |
Properties
| Property | Summary |
|---|---|
int NumSplits { get; } |
|
string Description { get; } |
Methods
| Method | Summary |
|---|---|
DataSplitResult<T> Split(Matrix<T> X, Vector<T> y) |
|
IEnumerable<DataSplitResult<T>> GetSplits(Matrix<T> X, Vector<T> y) |