SMOTEENN<T>
Models & Types · AiDotNet.Preprocessing.ImbalancedLearning
Implements SMOTE combined with Edited Nearest Neighbors (SMOTEENN) for handling imbalanced datasets.
Implements: IResamplingStrategy<T>
For Beginners
This is a two-step process:
Step 1 (SMOTE): Create synthetic minority samples
- Increases minority class size
- Fills in the minority class region
Step 2 (ENN): Clean up the boundary
- Removes samples misclassified by their neighbors
- Cleans both majority AND synthetic minority samples
- Creates a cleaner decision boundary
Why combine them:
- SMOTE alone can create noisy synthetic samples
- ENN cleans up these noisy samples
- Result is a balanced AND clean dataset
Visual example:
Original: M M M M M M M M M m m
After SMOTE: M M M M M M M M M m m m m m m m
^ synthetic
After ENN: M M M M M M m m m m m m
^ ^
Noisy samples removed from both classes
References:
- Batista et al. (2004). “A Study of the Behavior of Several Methods for Balancing
Machine Learning Training Data”
How It Works
SMOTEENN first applies SMOTE to oversample the minority class, then applies ENN to remove noisy and borderline samples from both classes. This combination often produces better results than either method alone.
Constructors
| Constructor | Summary |
|---|---|
SMOTEENN(double samplingStrategy, int kNeighborsSMOTE, int kNeighborsENN, int? seed) |
Initializes a new instance of the SMOTEENN class. |
Properties
| Property | Summary |
|---|---|
string Name { get; } |
Gets the name of this resampling strategy. |
Methods
| Method | Summary |
|---|---|
ResamplingStatistics<T> GetStatistics() |
Gets statistics about the last resampling operation. |
ValueTuple<Matrix<T>, Vector<T>> Resample(Matrix<T> x, Vector<T> y) |
Resamples the dataset using SMOTE followed by ENN. |