v2

we have revised the code to explicitly describe the distance-based filtering mechanism applied during synthetic data generation.

ClusterDEBO: A Cluster-Assisted Differential Evolution-Based Hybrid Oversampling Algorithm

🚀 ClusterDEBO is a novel hybrid oversampling algorithm designed to address class imbalance problems in machine learning datasets. It integrates K-Means clustering with Differential Evolution (DE) to generate structured and diverse synthetic samples.

🔹 Key Features:

✅ Clustering-Based Approach: Uses K-Means to partition the minority class into meaningful clusters.
✅ Differential Evolution (DE): Generates synthetic samples within clusters using mutation and crossover operations.
✅ Noise Injection: Adds small variations to synthetic samples to improve diversity and prevent overfitting.
✅ Adaptive Sampling: Adjusts the number of generated samples based on the class imbalance ratio.

📂 Included Files:
• ClusterDEBO.py → The Python implementation of the ClusterDEBO algorithm.
• Example usage will be provided in future releases.

🛠️ Installation & Usage:

from ClusterDEBO import ClusterDEBO

Initialize ClusterDEBO

oversampler = ClusterDEBO(n_clusters=10, crossover_rate=0.6, F=0.3)

Apply oversampling

X_resampled, y_resampled = oversampler.sample(X, y)

📜 Citation:

If you use this code in your research, please consider citing it via the Zenodo DOI:
📌 [Insert DOI here once generated]

🔗 Repository & DOI:
• 📌 GitHub Repository: [Insert GitHub Repo Link]
• 📌 DOI (Zenodo): [Insert Zenodo DOI]

🛠️ Contributions & Issues:
Feel free to open an issue or submit a pull request for improvements! 🚀