DEB-ADA: DENGESİZ VERİLERLE BAŞA ÇIKMAK İÇİN HİBRİT BİR YAKLAŞIM

Abstract

Denetimli öğrenme alanındaki sınıflandırma algoritmaları, sınıflar arası örnek sayılarının birbirine yakın olduğu dengeli verilerde en verimli şekilde çalışacak şekilde tasarlanmıştır. Bununla birlikte, gerçek dünya uygulamalarına bakıldığında, bazı sınıfların örnek sayıları diğer sınıflara kıyasla çok az veya çok fazla olabilir. Bu dengesizlik, sınıflandırma için eğitilmiş modellerin güvenilmez sonuçlar üretmesine neden olabilir. Bu çalışmada, dengesiz verilerle başa çıkmak için DEB ADA adı verilen yeni bir hibrit yaklaşım önerilmiştir. Yöntem, ADASYN ve DEBOHID algoritmalarının ardışık uygulanmasına dayanmaktadır. Hibrit yaklaşımda kullanılan ADASYN yöntemi, yüksek zorluk derecesine sahip azınlık sınıfı örneklerinin etrafında daha fazla sentetik örnek üreterek karar sınırlarını iyileştirir. Ayrıca, önerilen yaklaşımda DEBOHID yönteminin kullanılması aşırı uyum riskini azaltır. Önerilen yaklaşımın başarısı, KEEL veri deposundan alınan 44 dengesiz veri kümesi üzerinde test edilmiş ve Karar Ağacı (DT) sınıflandırıcı kullanılarak AUC ve G ortalama metrikleri ile değerlendirilmiştir. Deneysel sonuçlar, DEB ADA'nın literatürde yaygın olarak kullanılan klasik ve hibrit yöntemlere kıyasla daha tutarlı ve üstün sonuçlar verdiğini göstermektedir. Elde edilen bulgular, önerilen yöntemin dengesiz veri sınıflandırma problemine sağlam ve güvenilir bir çözüm sunduğunu göstermektedir.

Keywords

• Aşırı Örnekleme
• Az Örnekleme
• Dengesiz Öğrenme
• ADASYN
• DEBOHID

Supporting Institution

Bu çalışmanın hazırlanması aşamasında herhangi bir kurumdan destek alınmamıştır.

Ethical Statement

Bu çalışmanın hazırlık sürecinde bilimsel ve etik ilkelere uyulduğunu ve yararlanılan tüm çalışmaların kaynakçada belirtildiğini beyan ederim.

DEB-ADA: A HYBRID APPROACH TO DEALING WITH IMBALANCED DATA

Abstract

Classification algorithms in the field of supervised learning are designed to operate most efficiently on balanced data where the sample numbers across classes are close. However, when examining real-world applications, the sample numbers for some classes can be very few or very many compared to other classes. This imbalance can cause models trained for classification to produce unreliable results. In this study, a novel hybrid approach called DEB-ADA is proposed to deal with imbalanced data. The method is based on the sequential implementation of the ADASYN and DEBOHID algorithms. The ADASYN method used in the hybrid approach improves decision boundaries by generating more synthetic samples around the minority class examples with high difficulty. Furthermore, the use of the DEBOHID method in the proposed approach reduces the risk of overfitting. The success of the proposed approach was tested on 44 imbalanced datasets from the KEEL data repository and evaluated using the AUC and G-mean metrics using a Decision Tree (DT) classifier. Experimental results show that DEB-ADA delivers more consistent and superior results compared to classical and hybrid methods commonly used in literature. The findings demonstrate that the proposed method offers a robust and reliable solution to the imbalanced data classification problem.

Keywords

• Oversampling
• Undersampling
• Imbalanced Learning
• ADASYN
• DEBOHID

Supporting Institution

No support was received from any institution during the preparation of this study.

Ethical Statement

It is declared that during the preparation process of this study, scientific and ethical principles were followed, and all the studies benefited from are stated in the bibliography.

References

1. Arabie, P., Hubert, L., & De Soete, G. (1996). Clustering and classification. World Scientific.
2. Bach, M., Werner, A., & Palt, M. (2019). The proposal of undersampling method for learning from imbalanced datasets. Procedia Computer Science, 159, 125-134.
3. Batista, G. E., Prati, R. C., & Monard, M. C. (2004). A study of the behavior of several methods for balancing machine learning training data. ACM SIGKDD explorations newsletter, 6(1), 20-29.
4. Blagus, R., & Lusa, L. (2013). SMOTE for high-dimensional class-imbalanced data. BMC bioinformatics, 14(1), 106.
5. Boddu, A. S., & Jan, A. (2025). A systematic review of machine learning algorithms for breast cancer detection. Tissue and Cell, 102929.
6. Bougaham, A., El Adoui, M., Linden, I., & Frénay, B. (2024). Composite score for anomaly detection in imbalanced real-world industrial dataset. Machine Learning, 113(7), 4381-4406.
7. Bunkhumpornpat, C., Sinapiromsaran, K., & Lursinsap, C. (2009). Safe-level-smote: Safe-level-synthetic minority over-sampling technique for handling the class imbalanced problem. Pacific-Asia conference on knowledge discovery and data mining.
8. Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research, 16, 321-357.

9. Chen, W., Yang, K., Yu, Z., Shi, Y., & Chen, C. P. (2024). A survey on imbalanced learning: latest research, applications and future directions. Artificial Intelligence Review, 57(6), 137.
10. Chen, Z., Duan, J., Kang, L., & Qiu, G. (2021). A hybrid data-level ensemble to enable learning from highly imbalanced dataset. Information Sciences, 554, 157-176.
11. Cinar, A. C. (2025). A synergistic oversampling technique with differential evolution and safe level synthetic minority oversampling. Applied Soft Computing, 172, 112819.
12. Das, R., Biswas, S. K., Devi, D., & Sarma, B. (2020). An oversampling technique by integrating reverse nearest neighbor in SMOTE: Reverse-SMOTE. 2020 international conference on smart electronics and communication (ICOSEC).
13. Derrac, J., Garcia, S., Sanchez, L., & Herrera, F. (2015). Keel data-mining software tool: Data set repository, integration of algorithms and experimental analysis framework. J. Mult. Valued Logic Soft Comput, 17, 255-287.
14. Desuky, A. S., & Hussain, S. (2021). An improved hybrid approach for handling class imbalance problem. Arabian journal for science and engineering, 46(4), 3853-3864.
15. Fernández, A., Garcia, S., Herrera, F., & Chawla, N. V. (2018). SMOTE for learning from imbalanced data: progress and challenges, marking the 15-year anniversary. Journal of artificial intelligence research, 61, 863-905.
16. Gazzah, S., Hechkel, A., & Amara, N. E. B. (2015). A hybrid sampling method for imbalanced data. 2015 IEEE 12th International Multi-Conference on Systems, Signals & Devices (SSD15).
17. Gu, Q., Cai, Z., Zhu, L., & Huang, B. (2008). Data mining on imbalanced data sets. 2008 International Conference on advanced computer theory and engineering.
18. Han, H., Wang, W.-Y., & Mao, B.-H. (2005). Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning. International conference on intelligent computing.
19. He, H., Bai, Y., Garcia, E. A., & Li, S. (2008). ADASYN: Adaptive synthetic sampling approach for imbalanced learning. 2008 IEEE international joint conference on neural networks (IEEE world congress on computational intelligence).
20. Imani, M., Beikmohammadi, A., & Arabnia, H. R. (2025). Comprehensive analysis of random forest and XGBoost performance with SMOTE, ADASYN, and GNUS under varying imbalance levels. Technologies, 13(3), 88.
21. Karabiyik, M. A., Turkoglu, B., & Asuroglu, T. (2025). A cluster-assisted differential evolution-based hybrid oversampling method for imbalanced datasets. PeerJ Computer Science, 11, e3177.
22. Kaya, E., Korkmaz, S., Sahman, M. A., & Cinar, A. C. (2021). DEBOHID: A differential evolution based oversampling approach for highly imbalanced datasets. Expert Systems with Applications, 169, 114482.
23. Korkmaz, S. (2025). Hybridization of DEBOHID with ENN algorithm for highly imbalanced datasets. Engineering Science and Technology, an International Journal, 63, 101976.
24. Korkmaz, S., Şahman, M. A., Cinar, A. C., & Kaya, E. (2021). Boosting the oversampling methods based on differential evolution strategies for imbalanced learning. Applied Soft Computing, 112, 107787.
25. Kovács, G. (2019). Smote-variants: A python implementation of 85 minority oversampling techniques. Neurocomputing, 366, 352-354.
26. Liao, J.-J., Shih, C.-H., Chen, T.-F., & Hsu, M.-F. (2014). An ensemble-based model for two-class imbalanced financial problem. Economic Modelling, 37, 175-183.
27. Liu, D., Zhong, S., Lin, L., Zhao, M., Fu, X., & Liu, X. (2023). Deep attention SMOTE: Data augmentation with a learnable interpolation factor for imbalanced anomaly detection of gas turbines. Computers in Industry, 151, 103972.
28. Nikpour, B., Rahmati, F., Mirzaei, B., & Nezamabadi-pour, H. (2026). A comprehensive review on data-level methods for imbalanced data classification. Expert Systems with Applications, 295, 128920.
29. Öter, E., & Doğan, Y. (2024). A Comparative Study on Data Balancing Methods for Alzheimer's Disease Classification. Çukurova Üniversitesi Mühendislik Fakültesi Dergisi, 39(2), 489-501.
30. Özgün, S., & Şahman, M. A. (2024). Boosting the classification success in imbalanced data of bee larva cells. International Journal of Applied Methods in Electronics and Computers, 12(1), 10-15.
31. Pérez-Ortiz, M., Gutiérrez, P. A., Tino, P., & Hervás-Martínez, C. (2015). Oversampling the minority class in the feature space. IEEE transactions on neural networks and learning systems, 27(9), 1947-1961.
32. Ramentol, E., Caballero, Y., Bello, R., & Herrera, F. (2012). Smote-rs b*: a hybrid preprocessing approach based on oversampling and undersampling for high imbalanced data-sets using smote and rough sets theory. Knowledge and information systems, 33(2), 245-265.
33. Salekshahrezaee, Z., Leevy, J. L., & Khoshgoftaar, T. M. (2023). The effect of feature extraction and data sampling on credit card fraud detection. Journal of Big Data, 10(1), 6.
34. Santoso, B., Wijayanto, H., Notodiputro, K. A., & Sartono, B. (2017). Synthetic over sampling methods for handling class imbalanced problems: A review. IOP conference series: earth and environmental science.
35. Siddavatam, K. I., & Shinde, S. K. (2025). A Hybrid Literature Review on Handling Imbalanced Medical Data: AI Models and Open Issues. Expert Systems with Applications, 129004.
36. Spelmen, V. S., & Porkodi, R. (2018). A review on handling imbalanced data. 2018 international conference on current trends towards converging technologies (ICCTCT).
37. Storn, R., & Price, K. (1997). Differential evolution–a simple and efficient heuristic for global optimization over continuous spaces. Journal of global optimization, 11(4), 341-359. ŞAHMAN, M. A. (2021). Parameter analysis of differential evolution based oversampling approach for highly imbalanced datasets. International Journal of Intelligent Systems and Applications in Engineering, 9(2), 69-80.
38. Wheelus, C., Bou-Harb, E., & Zhu, X. (2018). Tackling class imbalance in cyber security datasets. 2018 IEEE International Conference on Information Reuse and Integration (IRI).
39. Wong, T.-T., & Chung, P.-C. (2025). A consistency analysis on four evaluation metrics for classifying imbalanced data. Knowledge and information systems, 1-18.
40. Xu, Z., Shen, D., Nie, T., Kou, Y., Yin, N., & Han, X. (2021). A cluster-based oversampling algorithm combining SMOTE and k-means for imbalanced medical data. Information Sciences, 572, 574-589.
41. Yildizdan, G., & Baykan, Ö. K. (2020). A novel modified bat algorithm hybridizing by differential evolution algorithm. Expert Systems with Applications, 141, 112949.