zlibsvm

zlibsvm is an object-oriented Java binding for the LIBSVM library. It wraps the cross-compiled Java code behind a clean API that can be easily integrated via Apache Maven.

Requirements

• Java 21+
• Maven 3.9.0+

Maven Dependency

<dependency>
    <groupId>de.hs-heilbronn.mi</groupId>
    <artifactId>zlibsvm-core</artifactId>
    <version>3.0.0</version>
</dependency>

Quickstart

A UI-based example project is available at rzo1/zlibsvm-example.

Dataset Format

zlibsvm uses the standard LIBSVM data format:

label feature_id1:feature_value1 feature_id2:feature_value2 ...

Each feature needs a unique integer identifier. For example, given three classes 1, 2, 3 and features a(id=1), b(id=2), c(id=3):

2 1:0.5325 3:0.523
3 2:0.7853 3:0.6326
1 1:0.53265 2:0.5422

Features not present in a data point can simply be omitted.

Implementing SvmDocument and SvmFeature

You need to provide implementations of SvmDocument and SvmFeature:

public class SvmDocumentImpl implements SvmDocument {

    private final List<SvmFeature> features;
    private final List<SvmClassLabel> classLabels = new ArrayList<>();

    public SvmDocumentImpl(List<SvmFeature> features) {
        this.features = features;
    }

    public List<SvmFeature> getSvmFeatures() {
        return features;
    }

    public SvmClassLabel getClassLabelWithHighestProbability() {
        if (classLabels.isEmpty()) {
            return null;
        }
        return Collections.max(classLabels);
    }

    @Override
    public List<SvmClassLabel> getAllClassLabels() {
        return Collections.unmodifiableList(classLabels);
    }

    @Override
    public void addClassLabel(SvmClassLabel classLabel) {
        assert (classLabel != null);
        this.classLabels.add(classLabel);
    }
}

public record SvmFeatureImpl(int index, double value) implements SvmFeature {

    public int getIndex() {
        return index;
    }

    public double getValue() {
        return value;
    }

    @Override
    public int compareTo(SvmFeature o) {
        return Integer.compare(getIndex(), o.getIndex());
    }
}

Training

Create a trainer with an SvmConfigurationImpl.Builder and train a model. The default configuration matches LIBSVM defaults: C-SVC with RBF kernel, gamma = 0, cost = 1.

SvmTrainer trainer = new SvmTrainerImpl(
        new SvmConfigurationImpl.Builder().build(), "my-model");

SvmModel model = trainer.train(documentsForTraining);

Prediction

Use the trained model to classify new documents:

SvmClassifier classifier = new SvmClassifierImpl(model);

List<SvmDocument> classified = classifier.classify(documentsForPrediction, true);

for (SvmDocument d : classified) {
    System.out.println(d + " was classified as category: "
            + d.getClassLabelWithHighestProbability().getNumeric());
}

License

Published under Apache License 2.0