MaximumDetector<T>

Models & Types · AiDotNet.AnomalyDetection.Ensemble

Combines multiple anomaly detectors using maximum score strategy.

Inheritance: ModelBase<T, Matrix<T>, Vector<T>> → AnomalyDetectorBase<T> → MaximumDetector<T>

Implements: IGeneratedParameterRegistrar<T>

For Beginners

This ensemble method takes the maximum anomaly score across all base detectors. If ANY detector thinks a point is anomalous, it gets a high score. This is useful when you want to catch anomalies that only specific detectors can find.

How It Works

The algorithm works by:

  1. Train each base detector on the data
  2. Normalize scores from each detector to [0,1]
  3. Take the maximum normalized score for each point
  4. Points with high max score are anomalies

When to use:

  • When different anomaly types need different detectors
  • When you want high recall (catch most anomalies)
  • When false negatives are costly

Constructors

Constructor Summary
MaximumDetector(double contamination, int randomSeed) Creates a new Maximum ensemble anomaly detector.

Methods

Method Summary
Vector<T> ScoreAnomalies(Matrix<T> X)
void Fit(Matrix<T> X)