NGBoostClassifier<T>

Models & Types · AiDotNet.Classification.Boosting

NGBoost (Natural Gradient Boosting) classifier for probabilistic classification.

Inheritance: ClassifierBase<T> → ProbabilisticClassifierBase<T> → EnsembleClassifierBase<T> → NGBoostClassifier<T>

For Beginners

Traditional classifiers give you a class prediction like “this email is spam” with maybe a confidence score. But NGBoost gives you properly calibrated probabilities - if it says “70% spam”, then about 70% of similar predictions will actually be spam.

Key benefits:

  • Well-calibrated probability estimates
  • Quantifies prediction uncertainty
  • Uses natural gradients for stable, efficient learning
  • Works well for both binary and multi-class problems

How It Works

NGBoost is an algorithm for probabilistic prediction that uses natural gradients to efficiently and directly optimize a proper scoring rule. For classification, it predicts class probabilities that are well-calibrated.

Reference: Duan, T., et al. “NGBoost: Natural Gradient Boosting for Probabilistic Prediction” (2019). https://arxiv.org/abs/1910.03225

Example

// Create NGBoost classifier for well-calibrated probabilistic predictions
var options = new NGBoostClassifierOptions<double>();
            
// Prepare training data: 6 samples with 2 features
var features = new Matrix<double>(6, 2);
features[0, 0] = 1.0; features[0, 1] = 1.1;
features[1, 0] = 1.2; features[1, 1] = 0.9;
features[2, 0] = 0.8; features[2, 1] = 1.0;
features[3, 0] = 5.0; features[3, 1] = 5.1;
features[4, 0] = 5.2; features[4, 1] = 4.9;
features[5, 0] = 4.8; features[5, 1] = 5.0;
var labels = new Vector<double>(new double[] { 0, 0, 0, 1, 1, 1 });
            
// Train using natural gradient boosting for calibrated probabilities
var result = new AiModelBuilder<double, Matrix<double>, Vector<double>>()
    .ConfigureModel(new NGBoostClassifier<double>(options))
    .Build(features, labels);
            
// Predict with calibrated probability estimates
var newSample = new Matrix<double>(1, 2);
newSample[0, 0] = 1.1; newSample[0, 1] = 1.0;
var prediction = result.Predict(newSample);
// Result is available in the returned value

Constructors

Constructor Summary
NGBoostClassifier(NGBoostClassifierOptions<T> options, IRegularization<T, Matrix<T>, Vector<T>> regularization) Initializes a new instance of NGBoostClassifier.

Properties

Property Summary
int NumberOfTrees { get; } Gets the number of trees in the ensemble.

Methods

Method Summary
Matrix<T> PredictLogProbabilities(Matrix<T> input) Predicts log probabilities for the input samples.
Matrix<T> PredictProbabilities(Matrix<T> input) Predicts class probabilities for the input samples.
ModelMetadata<T> GetModelMetadata()
Vector<T> Predict(Matrix<T> input) Predicts class labels for the input samples.
void Train(Matrix<T> x, Vector<T> y) Returns the model type identifier for this classifier.