RAkELClassifier<T>
Models & Types · AiDotNet.Classification.MultiLabel
Implements RAkEL (Random k-Labelsets) for multi-label classification.
Inheritance: MultiLabelClassifierBase<T> → RAkELClassifier<T>
For Beginners
RAkEL solves multi-label classification by training multiple Label Powerset classifiers on random subsets of k labels each. This captures label correlations (like Label Powerset) while avoiding the exponential explosion of label combinations.
How It Works
How it works:
- Randomly partition labels into overlapping subsets of size k
- Train a Label Powerset classifier on each subset
- Combine predictions by voting across all classifiers that predict each label
Key parameters:
- k: Size of each labelset (default: 3). Larger k captures more correlations but has more classes.
- numLabelsets: Number of random labelsets to create (default: 2*numLabels)
Reference: Tsoumakas et al., “Random k-Labelsets for Multilabel Classification” (2011)
Example
var features = new Matrix<double>(new double[,] { { 1.0, 2.0 }, { 3.0, 4.0 }, { 5.0, 6.0 }, { 7.0, 8.0 } });
var labels = new Matrix<double>(new double[,] { { 1.0, 0.0 }, { 0.0, 1.0 }, { 1.0, 1.0 }, { 0.0, 0.0 } });
var newSample = new Matrix<double>(new double[,] { { 1.0, 2.0 }, { 3.0, 4.0 }, { 5.0, 6.0 }, { 7.0, 8.0 } });
// Use AiModelBuilder facade for rakelclassifier classification
var builder = new AiModelBuilder<double, Matrix<double>, Matrix<double>>()
.ConfigureModel(new RAkELClassifier<double>(labelsetSize: 3));
var result = builder.Build(features, labels);
var prediction = result.Predict(newSample);
Constructors
| Constructor | Summary |
|---|---|
RAkELClassifier(int labelsetSize, int numLabelsets, int? seed, ClassifierOptions<T> options, IRegularization<T, Matrix<T>, Vector<T>> regularization) |
Creates a new RAkEL classifier. |
Properties
| Property | Summary |
|---|---|
int LabelsetSize { get; set; } |
Gets the size of each labelset (k parameter). |
int NumLabelsets { get; set; } |
Gets the number of labelsets to create. |
Methods
| Method | Summary |
|---|---|
Matrix<T> PredictMultiLabelProbabilities(Matrix<T> input) |
Predicts multi-label probabilities using RAkEL voting. |