기업부실 예측모형의 성과개선을 위한 기하평균 최적화 기반의 부스팅 학습 = Geometric Mean Optimization-based Boosting for Bankruptcy Prediction

국문 초록 (Abstract)

범주 불균형 문제는 분류 모형의 예측성과에 부정적인 영향을 미치는 문제로 이를 해결하기위한 다양한 접근 방법이 제안되어 왔다. 본 연구는 기업부실 예측문제에서 범주 불균형 문제에 적용된 AdaBoost 앙상블의 성과를 개선하기 위한 기하평균 최적화 기반의 부스팅 알고리즘인 GMOPTBoost를 제안한다. GMOPTBoost는 가우시안 경사하강법을 적용하여 기저 분류기에 할당된 최적의 가중치 집합을 탐색하여 예측 모형의 성과를 최적화한다. 범주불균형이 AdaBoost 앙상블 학습의 예측성과에 미치는 효과와 GMOPTBoost의 성과 개선효과를 검증하기 위하여 불균형 비율이 상이한 4개의 표본 군을 구성하여 10-fold 교차타당성 검증을 3회 수행한 결과는 다음과 같다. 첫째, 범주 불균형 문제는 예측성과에 부정적인영향을 미친다. 불균형 비율이 증가할수록 모형의 예측 성과가 감소한다. 둘째, GMOPTBoost는 불균형 데이터에 적용된 AdaBoost의 성과를 1% 유의수준에서 개선시키는 긍정적인 효과를 제공한다. 셋째, AdaBoost의 성과와 비교하여 GMOPTBoost는 보다 안정적인 예측 성과를 보여주었다

참고문헌 (Reference)

1 G. Zhang, "Theory and Methodology Artificial neural networks in bankruptcy prediction: General frame work and cross-validation analysis" 116 (116): 16-32, 1999

2 N. Japkowicz, "The class imbalance problem: A systematic study" 6 (6): 2002

3 Q. YU, "The Performance Stability of Defect Prediction Models with Class Imbalance: An Empirical Study" 100 (100): 2017

4 N. V. Chawla, "SMOTEBoost: Improving Prediction of the Minority Class in Boosting" 107-119, 2003

5 Y. Qu, "Review of bankruptcy prediction using machine learning and deep learning techniques" 162 : 895-899, 2019

6 C. Seiffert, "RUSBoost: Improving classification performance when training data is skewed" 2008

7 Soo Y. Kim, "Predicting restaurant financial distress using decision tree and AdaBoosted decision tree models" 36 : 354-362, 2014

8 A. Somasundaram, "Parallel and incremental credit card fraud detection model to handle concept drift and data imbalance" 31 (31): 2019

9 T. Le, "Over sampling techniques for bankruptcy prediction: Novel features from a transaction dataset" 10 (10): 2018

10 Young S. KWON, "Ordinal Pairwise Partitioning (OPP) Approach to Neural Networks Training in Bond rating" 6 (6): 1997

11 F. Barboza, "Machine learning models and bankruptcy prediction" 83 : 405-417, 2017

12 Financial Security Institute, "Machine learning adoption in domestic and foreign financial institutions"

13 M. Kubat, "Machine Learning for the Detection of Oil Spills in Satellite Radar Images" 30 : 195-215, 1998

14 S. Barua, "MWMOTE-Majority Weighted Minority Oversampling Technique for Imbalanced Data Set Learning" 26 (26): 405-425, 2014

15 Chosun Biz, "Loan review by analyzing non-financial information with AI"

16 W. F. Messier, "Inducing Rules for Expert System Development: An Example Using Default and Bankruptcy Data" 34 (34): 1988

17 K. Li, "Improved PSO_AdaBoost ensemble algorithm for imbalanced data" 19 (19): 2019

18 Myoung Jong Kim, "Geometric mean based boosting algorithm with over-sampling to resolve data imbalance problem for bankruptcy prediction" 42 (42): 08-, 2015

19 E. I. Altman, "Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy" 23 (23): 1968

20 J. A. Ohlson, "Financial Ratios and the Probabilistic Prediction of Bankruptcy" 18 (18): 109-131, 1980

21 A. Mellor, "Exploring issues of training data imbalance and mislabelling on random forest performance for large area land cover classification using the ensemble margin" 105 : 155-168, 2015

22 Myoung Jong Kim, "Ensemble with neural networks for bankruptcy prediction" 37 (37): 2010

23 M. Ziȩba, "Ensemble boosted trees with synthetic features generation in application to bankruptcy prediction" 58 : 93-101, 2016

24 W. Wei, "Effective detection of sophisticated online banking fraud on extremely imbalanced data" 16 (16): 2013

25 Jae Seok Park, "Development background and major trends of Fintech" 33 (33): 52-58, 2016

26 W. C. Lin, "Clustering-based under sampling in class-imbalanced data" 409 (409): 17-26, 2017

27 M. Geng, "CNUSVM:Hybrid CNN-uneven SVM model for imbalanced visual learning" BigMM 186-193, 2016

28 P. Ravi Kumar, "Bankruptcy prediction in banks and firms via statistical and intelligent techniques – A review" 180 : 1-28, 2007

29 L. Nanni, "An experimental comparison of ensemble of classifiers for bankruptcy prediction and credit scoring" 36 (36): 3028-3033, 2009

30 Kyung Shik Shin, "An application of support vector machines in bankruptcy prediction model" 28 (28): 2005

31 Y. LI, "Adapting SVM for data sparseness and imbalance:a case study in information extraction" 15 (15): 2009

32 Chosun Biz, "AI credit rating by identifying 70,000variables"

33 M. D. Odom, "A neural network model for bankruptcy prediction" 2 : 163-168, 1990

34 T. Le, "A cluster-based boosting algorithm for bankruptcy prediction in a highly imbalanced dataset" 10 (10): 2018

35 Mi So Ahn, "A Trend of Big Data in the financial industry"

36 Won Young Cho, "A Study on the Changes of Competitive Environment and Fair Competition Policy in New Era of the 4th Industrial Revolution" Software Policy & Research Institute 2018