DetectorBasedFilter<T>

Models & Types ยท AiDotNet.Preprocessing.OutlierHandling

Wraps any IAnomalyDetector for use in a preprocessing pipeline.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> โ†’ DetectorBasedFilter<T>

For Beginners

This class bridges the gap between anomaly detection algorithms and data preprocessing pipelines. It lets you use any anomaly detector to clean your data before training a model.

How It Works

Usage Examples:

Integration with PreprocessingPipeline:

Constructors

Constructor Summary
DetectorBasedFilter(IAnomalyDetector<T> detector, FilterMode mode, int[] columnIndices) Creates a new detector-based filter for preprocessing.

Properties

Property Summary
FilterMode Mode { get; } Gets the filter mode (how anomalies are handled).
IAnomalyDetector<T> Detector { get; } Gets the underlying anomaly detector.
bool SupportsInverseTransform { get; } Gets whether this transformer supports inverse transformation.

Methods

Method Summary
Vector<T> GetAnomalyScores(Matrix<T> data) Gets the anomaly scores for each row in the data.
bool[] GetAnomalyMask(Matrix<T> data) Gets a mask indicating which rows were identified as anomalies.
string[] GetFeatureNamesOut(string[] inputFeatureNames) Gets the output feature names after transformation.