SmoteEnnAugmenter<T>
Models & Types · AiDotNet.Augmentation.Tabular
Implements SMOTE-ENN combination for imbalanced datasets.
Inheritance: AugmentationBase<T, Matrix<T>> → TabularAugmenterBase<T> → SmoteEnnAugmenter<T>
For Beginners
SMOTE-ENN combines SMOTE oversampling with Edited Nearest Neighbors (ENN) cleaning. ENN removes samples whose class differs from the majority of their k nearest neighbors, cleaning both majority and minority samples to improve class separation.
How It Works
How it works:
- Apply SMOTE to generate synthetic minority samples
- For each sample, find its k nearest neighbors
- If the majority of neighbors have a different class, remove the sample
Benefits over SMOTE-Tomek:
- More aggressive cleaning than Tomek links
- Removes both misclassified majority AND minority samples
- Better noise removal in overlapping regions
Reference: Batista et al., “A Study of the Behavior of Several Methods for Balancing Machine Learning Training Data” (2004)
Constructors
| Constructor | Summary |
|---|---|
SmoteEnnAugmenter(int smoteKNeighbors, int ennKNeighbors, double samplingRatio, double probability) |
Creates a new SMOTE-ENN augmenter. |
Properties
| Property | Summary |
|---|---|
double SamplingRatio { get; } |
Gets the SMOTE sampling ratio. |
int EnnKNeighbors { get; } |
Gets the number of neighbors used by ENN. |
int SmoteKNeighbors { get; } |
Gets the number of nearest neighbors used by SMOTE. |
Methods
| Method | Summary |
|---|---|
IDictionary<string, object> GetParameters() |
|
ValueTuple<Matrix<T>, Vector<T>> ApplySmoteEnn(Matrix<T> data, Vector<T> labels, AugmentationContext<T> context) |
Applies SMOTE-ENN to a labeled dataset. |