Information Systems Review (경영정보학연구)

The Korea Society of Management Information Systems (한국경영정보학회)
권호 표지
DOI QR코드

DOI QR Code

Optimization of Uneven Margin SVM to Solve Class Imbalance in Bankruptcy Prediction

비대칭 마진 SVM 최적화 모델을 이용한 기업부실 예측모형의 범주 불균형 문제 해결

  • Received : 2022.08.04
  • Accepted : 2022.10.28
  • Published : 2022.11.30
Download PDF
⟨ Previous Next ⟩

Abstract

Although Support Vector Machine(SVM) has been used in various fields such as bankruptcy prediction model, the hyperplane learned by SVM in class imbalance problem can be severely skewed toward minority class and has a negative impact on performance because the area of majority class is expanded while the area of minority class is invaded. This study proposed optimized uneven margin SVM(OPT-UMSVM) combining threshold moving or post scaling method with UMSVM to cope with the limitation of the traditional even margin SVM(EMSVM) in class imbalance problem. OPT-UMSVM readjusted the skewed hyperplane to the majority class and had better generation ability than EMSVM improving the sensitivity of minority class and calculating the optimized performance. To validate OPT-UMSVM, 10-fold cross validations were performed on five sub-datasets with different imbalance ratio values. Empirical results showed two main findings. First, UMSVM had a weak effect on improving the performance of EMSVM in balanced datasets, but it greatly outperformed EMSVM in severely imbalanced datasets. Second, compared to EMSVM and conventional UMSVM, OPT-UMSVM had better performance in both balanced and imbalanced datasets and showed a significant difference performance especially in severely imbalanced datasets.

Support Vector Machine(SVM)은 기업부실 예측문제 등 다양한 분야에서 성공적으로 활용되어 왔으나 범주 불균형 문제가 존재하는 경우 다수 범주의 경계영역은 확장되는 반면, 소수 범주의 경계영역은 축소되고 분류 경계선이 소수 범주로 편향되어 분류 성과에 부정적인 영향을 미치는 것으로 보고되고 있다. 본 연구는 범주 불균형 문제에 대한 대칭 마진 SVM(EMSVM)의 한계점을 개선하기 위하여 비대칭 마진 SVM(UMSVM)과 임계점 이동 기법을 결합한 최적화 비대칭 마진 SVM인 OPT-UMSVM을 제안한다. OPT-UMSVM은 소수 범주 방향으로 치우진 분류 경계선을 다수 범주로 재이동함으로써 소수 범주의 민감도를 개선하고 최적화된 분류 성과를 산출함으로써 SVM의 일반화 능력을 향상시키는 장점을 가진다. OPT-UMSVM의 성과 개선 효과를 검증하기 위하여 불균형 비율이 상이한 5개의 표본군을 구성하여 10-fold 교차타당성 검증을 수행한 결과는 다음과 같다. 첫째, 범주 불균형이 미미한 표본에서 UMSVM은 EMSVM의 성과 개선 효과가 미약한 반면, 범주 불균형이 심화된 표본에서 UMSVM은 EMSVM의 성과개선에 크게 공헌하고 있다. 둘째, OPT-UMSVM은 EMSVM 및 기존의 UMSVM과 비교하여 범주 균형 및 범주 불균형 표본 모두에서 보다 우수한 성과를 가지고 있으며, 특히 범주 불균형이 심화된 표본에서 유의적인 성과 차이를 보였다.

Keywords

Acknowledgement

이 논문은 2022년도 정부(과학기술정보통신부)의 재원으로 정보통신기획평가원의 지원을 받아 수행된 연구임(IITP-2022-0-01201).

References

  1. 박종원, 안성만, "재무비율을 이용한 부도예측에 대한 연구: 한국의 외부감사대상기업을 대상으로", 경영학연구, 제43권, 제3호, 2014, pp. 639-669
  2. 윤우섭, 김명종, "AUROC기반의 부도예측 앙상블 모형", 중소기업금융연구, 2021, 제41권 제3호 pp. 41-60 https://doi.org/10.33219/JSMEF.2021.41.3.002
  3. Abd Elrahman, S. M. and A. Abraham, "A review of class imbalance problem", Network and Innovative Computing, Vol.1, 2013, pp. 332-340.
  4. Altman, E. I., "Financial ratios, discriminant analysis and the prediction of corporate bankruptcy", The Journal of Finance, Vol.23, No.4, 1968, pp. 589-609. https://doi.org/10.1111/j.1540-6261.1968.tb00843.x
  5. Barboza, F., H. Kimura, and E. Altman, "Machine learning models and bankruptcy prediction", Expert Systems with Applications, Vol.83, 2017, pp. 405-417. https://doi.org/10.1016/j.eswa.2017.04.006
  6. Beaver, W., "Financial ratios as predictors of failure", Journal of Accounting Research, Vol.71, No.4, 1966, pp. 71-111. https://doi.org/10.2307/2490171
  7. Burges, C. J. C., "A tutorial on support vector machines for pattern recognition", Data Mining and Knowledge Discovery, Vol.2, 1998, pp. 121-167. https://doi.org/10.1023/A:1009715923555
  8. Cao, J., H. Lu, W. Wang, and J. Wang, "A loan default discrimination model using cost-sensitive support vector machine improved by PSO", Information Technology and Management, Vol.14, No.3, 2013, pp. 193-204. https://doi.org/10.1007/s10799-013-0161-1
  9. Chen, M. Y., "Bankruptcy prediction in firms with statistical and intelligent techniques and a comparison of evolutionary computation approaches", Computers and Mathematics with Applications, Vol.62, No.12, 2011, pp. 4514-4524. https://doi.org/10.1016/j.camwa.2011.10.030
  10. Cortes, C., V. Vapnik, and L. Saitta, Support- Vector Networks Editor, Machine Learning, Kluwer Academic Publishers, 1995, pp. 273-297.
  11. Dal Pozzolo, A., O. Caelen, Y. A. le Borgne, S. Waterschoot, and G. Bontempi, "Learned lessons in credit card fraud detection from a practitioner perspective", Expert Systems with Applications, Vol.41, No.10, 2014, pp. 4915-4928. https://doi.org/10.1016/j.eswa.2014.02.026
  12. Davis, J. and M. Goadrich, "The relationship between Precision-Recall and ROC curves", Proceedings of the 23rd International Conference on Machine Learning - ICML '06. 2, 2006, pp. 33-240.
  13. Du, J., Vong, C. M., C. M. Pun, P. K. Wong, and W. F. Ip, "Post-boosting of classification boundary for imbalanced data using geometric mean", Neural Networks, Vol.96, 2017, pp. 101-114. https://doi.org/10.1016/j.neunet.2017.09.004
  14. Fawcett, T., "An introduction to ROC analysis", Pattern Recognition Letters, Vol.27, No.8, 2006, pp. 861-874. https://doi.org/10.1016/j.patrec.2005.10.010
  15. Feng, W., W. Huang, and J. Ren, "Class imbalance ensemble learning based on the margin theory", Applied Sciences, Switzerland, Vol.8, No.5, 2018, p. 815.
  16. Galar, M., A. Fernandez, E. Barrenechea, H. Bustince, and F. Herrera, "A review on ensembles for the class imbalance problem: Bagging-, boosting-, and hybrid-based approaches", IEEE Transactions on Systems, Man and Cybernetics Part C: Applications and Reviews, Vol.42, No.4, 2012, pp. 463-484
  17. Geng, M., Y. Wang, Y. Tian, and T. Huang, "CNUSVM: Hybrid CNN-uneven SVM model for imbalanced visual learning", Proceedings - 2016 IEEE 2nd International Conference on Multimedia Big Data, BigMM 2016, 2016, pp. 186-193.
  18. He, H. and E. A. Garcia, "Learning from imbalanced data", IEEE Transactions on Knowledge and Data Engineering, Vol.21, No.9, 2009, pp. 1263-1284. https://doi.org/10.1109/TKDE.2008.239
  19. Horak, J., J. Vrbka, and P. Suler, "Support vector machine methods and artificial neural networks used for the development of bankruptcy prediction models and their comparison", Journal of Risk and Financial Management, Vol.13, No.3. 2020, p. 60.
  20. Kang, P. and S. Cho, "EUS SVMs: Ensemble of under-sampled SVMs for data imbalance problems", ICONIP 2006: Neural Information Processing, 2006, pp. 837-846.
  21. Kim, M. J., D. K. Kang, and H. B. Kim, "Geometric mean based boosting algorithm with over-sampling to resolve data imbalance problem for bankruptcy prediction", Expert Systems with Applications, Vol.42, No.3, 2015, pp. 1074-1082. https://doi.org/10.1016/j.eswa.2014.08.025
  22. Kim, T. and H. Ahn, "A hybrid under-sampling approach for better bankruptcy prediction", Journal of Intelligence and Information Systems, Vol.21, No.2, 2015, pp. 173-190. https://doi.org/10.13088/JIIS.2015.21.2.173
  23. Kubat, M., R. Holte, and S. Matwin, "Learning when negative examples abound", ECML 1997: Machine Learning: ECML-97, 1997, pp. 146-153.
  24. Kuspriyanto, S. O., D. H. Widyantoro, H. S. Sastramihardja, K. Muludi, and S. Maimunah, "Performance evaluation of SVM-based information extraction using τ margin values", International Journal on Electrical Engineering and Informatics, Vol.2, No.4, 2010, pp. 256-265. https://doi.org/10.15676/ijeei.2010.2.4.1
  25. Lee, H. K. and S. B. Kim, "An overlap-sensitive margin classifier for imbalanced and overlapping data", Expert Systems with Applications, Vol.98, 2018, pp. 72-83. https://doi.org/10.1016/j.eswa.2018.01.008
  26. Li, Y. and J. Shawe-Taylor, "The SVM With Uneven Margins and Chinese Document Categorisation", Language, Information and Computation : Proceedings of the 17th Pacific Asia Conference, Sentosa, Singapore, 2003, pp. 1-3.
  27. Li, Y., K. Bontcheva, and H. Cunningham, "Adapting SVM for data sparseness and imbalance: A case study in information extraction", Natural Language Engineering, Vol.15, No.2, 2009, pp. 241-271. https://doi.org/10.1017/S1351324908004968
  28. Li, Y., K. Bontcheva, and H. Cunningham, "Using Uneven Margins SVM and Perceptron for Information Extraction", 2005, Available at http://svmlight.joachims.org.
  29. Mazurowski, M. A., P. A. Habas, J. M. Zurada, J. Y. Lo, J. A. Baker, and G. D. Tourassi, "Training neural network classifiers for medical decision making: The effects of imbalanced datasets on classification performance", Neural Networks, Vol.21, No.2-3, 2008, pp. 427-436. https://doi.org/10.1016/j.neunet.2007.12.031
  30. Ni, W., T. Liu, J. Xu, Y. Huang, and J. Ge, "Acronym extraction using SVM with Uneven Margins", Proceedings - 2010 IEEE 2nd Symposium on Web Society, SWS 2010, 2010, pp. 132-138.
  31. Olson, D. L., D. Delen, and Y. Meng, "Comparative analysis of data mining methods for bankruptcy prediction", Decision Support Systems, Vol.52, No.2, 2012, pp. 464-473.
  32. Shawe-Taylor, J., "Classification accuracy based on observed margin", Algorithmica, Vol.22, 1998, pp. 157-172. https://doi.org/10.1007/PL00013827
  33. Sundarkumar, G. G. and V. Ravi, "A novel hybrid undersampling method for mining unbalanced datasets in banking and insurance", Engineering Applications of Artificial Intelligence, Vol.37, 2015, pp. 368-377. https://doi.org/10.1016/j.engappai.2014.09.019
  34. Tao, X., Li, Q., Guo, W., Ren, C., Li, C., Liu, R., & Zou, J., "Self-adaptive cost weights-based support vector machine cost-sensitive ensemble for imbalanced data classification", Information Sciences, Vol.487, 2019, pp. 31-56.
  35. Vapnik, V. N., "An overview of statistical learning theory", IEEE Transactions on Neural Networks, Vol.10, No.5, 1999, pp. 988-999. https://doi.org/10.1109/72.788640
  36. Veganzones, D. and E. Severin, "An investigation of bankruptcy prediction in imbalanced datasets", Decision Support Systems, Vol.112, 2018, pp. 111-124. https://doi.org/10.1016/j.dss.2018.06.011
  37. Veropoulos, K., C. Campbell, and N. Cristianini, "Controlling the Sensitivity of Support Vector Machines", Proceedings of the International Joint Conference on Artificial Intelligence, Stockhol, Sweden (IJCAI99), 1999, pp. 55-60.
  38. Wu, G. and E. Y. Chang, "KBA: Kernel boundary alignment considering imbalanced data distribution", IEEE Transactions on Knowledge and Data Engineering, Vol.17 No.6, 2005, pp. 786-795 https://doi.org/10.1109/TKDE.2005.95
  39. Yan, Q., S. Xia, and F. Meng, "Optimizing cost-sensitive SVM for imbalanced data: Connecting cluster to classification", arXiv:1702.01504, Cornell University, 2017, pp. 1-15.
  40. Zhang, T. and Z. H. Zhou, "Optimal margin distribution machine", IEEE Transactions on Knowledge and Data Engineering, Vol.32, No.6, 2020, pp. 1143-1156. https://doi.org/10.1109/TKDE.2019.2897662
  41. Zhou, L., "Performance of corporate bankruptcy prediction models on imbalanced dataset: The effect of sampling methods", Knowledge-Based Systems, Vol.41, 2013, pp. 16-25.
  42. Zou, Q., S. Xie, Z. Lin, M. Wu, and Y. Ju, "Finding the best classification threshold in imbalanced classification", Big Data Research, Vol.5, 2016, pp. 2-8. https://doi.org/10.1016/j.bdr.2015.12.001