Implementation:Scikit learn contrib Imbalanced learn SMOTETomek

Knowledge Sources	imbalanced-learn imbalanced-learn Docs
Domains	Machine_Learning, Data_Preprocessing, Imbalanced_Learning
Last Updated	2026-02-09 03:00 GMT

Overview

Concrete tool for combined SMOTE oversampling and Tomek Links cleaning provided by the imbalanced-learn library.

Description

The SMOTETomek class combines SMOTE oversampling with Tomek Links under-sampling. It first applies SMOTE to generate synthetic minority samples, then identifies and removes Tomek link pairs (nearest-neighbor pairs from different classes), cleaning the boundary. This is a gentler cleaning approach than SMOTEENN.

Usage

Import this class when you want combined oversampling and cleaning with minimal data removal. SMOTETomek removes only direct boundary ambiguities (Tomek links) rather than all misclassified neighbors.

Code Reference

Source Location

Repository: imbalanced-learn
File: imblearn/combine/_smote_tomek.py
Lines: L26-157

Signature

class SMOTETomek(BaseSampler):
    def __init__(
        self,
        *,
        sampling_strategy="auto",
        random_state=None,
        smote=None,
        tomek=None,
        n_jobs=None,
    ):
        """
        Args:
            sampling_strategy: str, dict, or callable - Resampling ratio.
            random_state: int, RandomState, or None - Seed.
            smote: SMOTE or None - SMOTE instance (default: SMOTE()).
            tomek: TomekLinks or None - TomekLinks instance (default: TomekLinks()).
            n_jobs: int or None - Parallel jobs.
        """

Import

from imblearn.combine import SMOTETomek

I/O Contract

Inputs

Name	Type	Required	Description
X	{array-like, sparse matrix} of shape (n_samples, n_features)	Yes	Feature matrix
y	array-like of shape (n_samples,)	Yes	Target labels
smote	SMOTE or None	No	Custom SMOTE instance
tomek	TomekLinks or None	No	Custom TomekLinks instance

Outputs

Name	Type	Description
X_resampled	ndarray of shape (n_samples_new, n_features)	Oversampled and Tomek-cleaned matrix
y_resampled	ndarray of shape (n_samples_new,)	Cleaned target array

Usage Examples

from collections import Counter
from sklearn.datasets import make_classification
from imblearn.combine import SMOTETomek

X, y = make_classification(
    n_classes=2, weights=[0.1, 0.9], n_samples=1000, random_state=10
)
smote_tomek = SMOTETomek(random_state=42)
X_res, y_res = smote_tomek.fit_resample(X, y)
print(f"Resampled: {Counter(y_res)}")

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment