Using GA based Input Selection Method for Artificial Neural Network Modeling Application to Bankruptcy Prediction

유전자 알고리즘을 활용한 인공신경망 모형 최적입력변수의 선정: 부도예측 모형을 중심으로

  • 홍승현 (이화여자대학교 경영학과) ;
  • 신경식 (이화여자대학교 경영학과)
  • Published : 2003.06.01

Abstract

Prediction of corporate failure using past financial data is a well-documented topic. Early studies of bankruptcy prediction used statistical techniques such as multiple discriminant analysis, logit and probit. Recently, however, numerous studies have demonstrated that artificial intelligence such as neural networks can be an alternative methodology for classification problems to which traditional statistical methods have long been applied. In building neural network model, the selection of independent and dependent variables should be approached with great care and should be treated as model construction process. Irrespective of the efficiency of a teaming procedure in terms of convergence, generalization and stability, the ultimate performance of the estimator will depend on the relevance of the selected input variables and the quality of the data used. Approaches developed in statistical methods such as correlation analysis and stepwise selection method are often very useful. These methods, however, may not be the optimal ones for the development of neural network model. In this paper, we propose a genetic algorithms approach to find an optimal or near optimal input variables fur neural network modeling. The proposed approach is demonstrated by applications to bankruptcy prediction modeling. Our experimental results show that this approach increases overall classification accuracy rate significantly.

부도예측모형의 구축은 은행 등 금융기관이 신용평가시스템 혹은 심사역 의사결정지원시스템을 구축하는데 중요한 기반이 된다. 많은 선행연구들에서는 기업의 부도예측을 위하여 전통적으로 다변량 판별분석이나 로짓분석과 같은 통계기법이 많이 사용되었으나, 최근에는 많은 연구들에 의해 그 우수성이 보고되고 있는 인공신경망, 귀납적 학습방법 등 인공지능 기법이 부도예측분야에 많이 응용되고 있다. 일반적으로 인공신경망 기법을 응용한 부도예측모형에서는 기업의 재무정보 및 비재무 정보를 입력변수로 주고 기업의 부도여부를 출력변수로 설정하여 학습을 통해 이들의 관계를 추출하고 있다. 그러므로 입력변수의 선정은 모형의 정확도에 커다란 영향을 미치며, 입력변수가 잘못 선정된 경우 예측 정확도는 현저히 낮아진다. 그러나 최적의 입력변수군을 선정하는 문제는 매우 어려운 과제 중 하나로, 선행 연구들에서는 주로 전문가의 의견을 반영하거나, 문헌을 통해 도출, 혹은 통계적 기법을 활용하여 입력변수를 선정하는 것이 일반적이었다. 본 연구에서는 많은 선행 연구에서 모형구축에의 한계점으로 명시하고 있는 입력변수 선정의 문제에 대해 유전자 알고리즘을 이용한 최적화를 통하여 입력 변수군을 도출하는 방법론을 제시하였고, 이 방법론이 다른 통계기법이나 전문가에 의한 변수 선택 방법론에 비해 우수함을 인공신경망 모형에 적용한 결과를 비교함으로 보여 주었으며, 이들간의 예측력의 차이가 유의함을 통계적 검증하였다. 모형의 실험을 위하여 총 528개사의 재무정보를 활용하였는데, 이는 1995년부터 1997년까지 3년간 부도가 발생한 일반법인 제조업체 중 외감법인 이상 264개사와 동수의 건전기업의 재무 데이터로 구성하였다. 기업이 도산에 이르기까지 많은 변인들이 다양하게 작용하게 된다. 그러나 이러한 변인들을 모두 모형에 적용하는 것은 비효율적이며, 인공신경망 모형에서 과다 입력변수를 사용하는 경우 수렴과 일반화 모두에 바람직하지 않은 결과가 나타난다. 따라서 적절한 입력변수군의 선택은 인공신경망 모형의 효율성과 성능을 향상시키게 되고, 이는 부도 예측율의 향상으로 이어질 수 있다. 이에 인공신경망 모형을 위한 최적의 입력변수군을 선정하고자 한 본 연구는 결국 기업의 부도 예측율을 높이기 위한 방법론을 제시했다는 점에 그 의의가 있다.

Keywords

References

  1. 회계학연구 v.12 부가가치회계정보의 기업부실 예측력에 관한 실증적 연구 강철승
  2. 경영학연구 v.16 no.2 재무비율로 판단한 기업부실 징후와 예측 김건우
  3. 신경망 이론과 응용(Ⅰ) 김대수
  4. 메타 휴리스틱 김여근;윤복식;이상복
  5. 고려대학교 대학원 박사학위논문 우열기업예측을 위한 재무비율 선정에 관한 연구 김재권
  6. 서울대학교 대학원 박사학위논문 우리나라 재무제표 기업실패 예측능력에 관한 실증적 연구 박창길
  7. 한국경제 v.15 기업부실예측모델의 재정립을 통한 기업부실원인과의 연계에 대한 실증적 연구 송인만
  8. 한국경영정보학회 국제학술대회 논문집 Neuro-genetic Approach for Bankruptcy Prediction: A Comparison to Back-propagation Algorithms 신경식;신택수;한인구
  9. 한국경영과학회지 다수의 인공신경망을 통합한 기업부도 예측모형에 관한 연구 신경식;한인구
  10. 한국경영정보학회 국제학술대회 논문집 Corporate Failure Prediction Modeling Using Genetic Algorithm Technique 신경식;한인구
  11. 동국대학교 대학원 박사학위논문 지금흐름정보의 기업도산예측력에 관한 실증적 연구 윤주석
  12. 경영학연구 v.23 no.2 기업도산예측을 위한 귀납적 학습지원 인공신경망 접근방법 : MDA, 귀납적 학습방법, 인공신경망 모형과의 성과비교 이건창;김명종;김혁
  13. 한국경영과학회지 v.21 no.1 통계적 모형과 인공지능 모형을 결합한 기업신용평가 모형에 관한 연구 이건창;한인구;김명종
  14. 회계학연구 v.16 회계정보에 의한 기업부실예측과 시장반응 이계원
  15. 한국경영과학회 추계학술대회 발표논문집 인공신경망 학습단계에서의 Genetic Algorithm을 이용한 입력변수 선정 이재식;차봉근
  16. 한국전문가시스템학회지 v.1 no.1 인공신경망을 이용한 중소기업도산예측에 있어서의 비재무정보의 유용성 검증 이재식;한재홍
  17. 성균관대학교 대학원 박사학위논문 기업부실 예측에 관한 실증적 연구 임영규
  18. 기업도산예측모델 정준수
  19. 재무분석을 중심으로 한 기업의 부실화 예측기법연구 v.2 중소기업은행
  20. 기업도산의 예측에 관한 실증적 연구 지청
  21. 판별분석에 의한 기업평가방법 한국은행 은행감독원 여신관리국
  22. 한국경영과학회지 v.22 no.3 The hybrid systems for credit rating 한인구;조홍규;신경식
  23. 경희대학교 대학원 박사학위논문 기업도산 예측에 관한 연구 전춘옥
  24. 교려대학교 대학원 박사학위논문 재무제표 적시성에 관한 연구 허영빈
  25. 회계학연구 v.12 기업부실예측에 관한 연구 황석하
  26. Journal of Finance An Analysis of Risk and Return Characteristics of Corporate Bankruptcy using Capital Market Data Aharony,J.;Jonew,C.P.;Swary,I.
  27. The Journal of Finance Financial ratios, discriminant analysis and the prediction of corporate bankruptcy Altman,E.I.
  28. Journal of Banking and Finance ZETA Analysis: A New Model to Identify Bankruptcy Risk of Corporation Altman,E.I.;Haldeman,R.G.;Narayanan,P.
  29. Financial Management Cash Folw Reporting and Financial Distress Models: Testing of Hypotheses Aziz,A.;Lawson,G.H.
  30. Proceedings of The third World Congress on Expert Systems Neural Networks and Genetic Algorithms for Bankruptcy Predictions Barbro,B.;Teija L.;Kaisam S.
  31. The Accounting Review Accounting Procedures, Market Data, Cash-flow Figures, and Insolvency Classification: The Case of the Insurance Industry Barniv,R.
  32. Intelligent Systems in Accounting, Finance and Management v.6 Predicting the outcome following bankruptcy filing: A three-state classification using neural networks Barniv, R.;Agawal, A.;Leach,R.
  33. Journal of Accounting Research v.5 Financial ratios as predictors of failure. Empirical Research in Accounting: Selected Studies Beaver,W.
  34. Intelligent Systems in Accounting, Finance and Management v.6 Neural nets or the logit model: A comparison of each model's ability to predict commercial bank failures Bell,T.
  35. Journal of Accounting Research v.12 no.1 Failing Company Discriminant Analysis Blum,M.
  36. Expert Systems with Applications v.9 Effectiveness of neural networks types for prediction of business failure Bortiz,J.;Kennedy,D.
  37. Proceedings of the First International Conference on Artificial Intelligence Application on Wall Street Neural Network and Mathematics of Chaos - An Investigation of These Methodologies as Accurate Predictors of Corporate Bankruptcy Cadden,D.T.
  38. The Accountion Review Discriminating between reorganized and liquidated firms in bankruptcy Casey,C.;McGee,V.;Stickney,C.
  39. Financial Management An Empirical Analysis of Useful Financial Ratios Chen,K.H.;Shimerda,T.A.
  40. Intelligent Systems in Accounting, Finance and Management v.2 A comparative analysis of inductive learning algorithm Chung,H.;Tam.K.
  41. The Journal of Finance The Behavior of the Common Stock of Bankruptcy Firms Clark,T.A.;Weinstein,M.I.
  42. Trading On The Edge Genetic algorithms for financial modeling Colin,A.M.;Deboeck,G.J.(Ed.)
  43. The Journal of Finance v.35 no.4 Ratio Stability and Corporate Failure Dambolena,I.G.;Khoury,S.I.
  44. Handbook of Genetic Algorithms Davis,L.
  45. Journal of Accounting Research v.10 A Discriminant Analysis of Predictor of Business Failure Deakin,E.B.
  46. Tanding On The Edge Using GAs to optimize a trading system Deboeck.G.J.;Deboeck,G.J.(Eds.)
  47. Journal of Finance and Quantitative Analysis An Empirical Test of Financial Ratio Analysis for Small Business Failure Prediction Edminster,R.O.
  48. The Accounting Review v.50 The Effect of Lease data on the predictive ability of financial ratios Elam,R.
  49. Financial Management An Expert System Approach to Financial Analysis: The Case of S&L Bankruptcy Elmer,P.J.;Borowski,D.M.
  50. Journal of Accounting Research v.20 no.2 Emery,G.W.;Cogger,K.O.
  51. Intellgent Systems In Accounting Finanace and Management v.6 A comparative of the relative costs of financial distress models: Arificial neural networks, logit and multivariate discriminant analysis Etheridge,H.;Sriram,R.
  52. The Certified Public Account v.2 A comparison of the ratios of successful industrial enterprises with those of failed companies Fitzpatrick.P.
  53. Information and Management v.24 no.3 Forecasting with neural networks: An application using bankruptcy data Fletcher,D.;Goss,E.
  54. The Accounting Review Quarterly Accounting Data: Time-Series Properties and Predictive-Ability Results Foster,G.
  55. Journal of Accounting Research Classifying Bankrupt Firms With Funds Flow Components Gentry,J.A.;Newbold,P.;Whitford,D.T.
  56. Genetic Algorithms in Search, Optimization and Machine Learning Glodberg,D.E.
  57. Financial Management Financial Ratio Patterns in Retail and Manufacturing Organizations Gombola M.J.;Ketz,J.E.
  58. Journal of Accounting and Public Pollicy Failure prediction: Sensitivity of Classification accuracy to alternative statistical method and variable sets Hamer,M.
  59. Journal of the Korean operations Research and Management Science Society v.22 no.3 The hybrid systems for credit rating Han,I.;Jo,H.;Shin,K.S.
  60. Research Papers in Banking and Economics, Financial Studies Section Predicting Bank Failure Hanweak,G.A.
  61. The Organization of Behavior: A Neuropsychological Theory Hebb.D.O.
  62. Neurocomputing Hecht-Nielsen,R.
  63. Adaptation in Natural and Artficial Systems Holland,J.H.
  64. Proceedings of PACES95 Corporate Analysis of Bankruptcy Prediction Accuracy: Using case-based forecasting, Neural network, and discriminant analysis Jo,H.;Han,I.;Lee,H.
  65. Journal of Financial and Quantitative Analysis v.14 The Cross-Sectional Stability of Financial Ratio Patterns Johnson,W.B.
  66. Journal of Accounting Research The Effect of General Price Level Adjustment on the Predictive Ability of Financial Ratios Ketz,J.E.
  67. Search Space Research Report no.1 Genetic algorithms for bankruptcy prediction Kingdom,J.;Feldman,K.
  68. Genetic programming Koza,J.
  69. Ph,D.thesis, The Wharton School, University of Pennsylvania Predictive insights through analogical reasoning Application to screening new financial service concepts Lee,H.Y.
  70. Proceedings of the 1994 IEEE International Conference on Neural Network Hybird Neural Network-Driven Reasoning Approach to Bankruptcy Prediction: Comparison with MDA, ACLS, and Neural Network Lee,K.C.;Kim,J.
  71. Expert Systems With Applications v.1 A comparative study of recursive partitioning algorithm and analogue concept learning system Lee,S.B.;Oh,S.H.
  72. Journal of Accountion Research Accounting Ratios and the Prediction of failure: Some Behavioral Evidence Libby,R.
  73. Expert Systems With Application v.13 no.2 Bankruptcy prediction using case-based reasoning, neural networks, and discriminant analysis Jo,H,;Han,I.;Lee,H.
  74. Proceedings of the 3rd International Conference on Artificial Intelligence Applications on Wall Street Genetic algorithms for predicting individual stock performance Mahfoud,S.;Mani,G.
  75. Journal of Accounting Research The Experimental Design of Classification models: An Application of Recursive Partitioning and Bootstrapping to Commercial Bank Loan Classifications Marais,M.;Patell,J.;Wolfson,M.
  76. Journal of Banking and Finance v.1 Early Warning of Bank Failure: A Logit Regression Approach Martin,D.
  77. Bulletin of Mathematical Biophysics v.5 A logical calclus of the ideas immanent in nervous activity McCulloch,W.S.;Pitts,W.
  78. European Journal of Operational Research v.138 Genetic programming and rough sets: A hybrid approach to bankruptcy classification McKee,T.E.;Lensberg T.
  79. The Accounting Review The Differential Bankruptcy Predictive Ability of Specific of Specific Price Level Adjustments: Some Empirical Evidence Mensah,Y.M.
  80. Journal of Accounting Research An Examination of the Stationarity of Multivariate Bankruptcy Prediction Models: A methodological study Mensah,Y.M.
  81. National Bureau of Research v.105 Financial Small Corporations in Five Manufacturing industries 1926-1936 Merwin,C.L.
  82. Proceedings of the Third International Conference on Artificial Intelligence Applications on Wall Street Bank Failure and Categorization- A Neural Network Approach Miller,W.;Cadden,D.T.;Driscoll,V.
  83. The Accounting Review A Comparison of General Price Level and Historical Cost Financial Statements in the Prediction of Bankruptcy Norton,C.L.;Smith,R.E.
  84. Proceedings of the IEEE International Conference on Neural Network v.2 A nerual networks model for bankrutcy prediction Odom,M.;Sharda,R.
  85. Journal of Accounting Research Financial ratios and the probablistic prediction of bankruptcy Ohlson,J.
  86. Proceedings of the Third International Conference on Artificial Intelligence Applications on Wall Street Optimal Mixtures of Classifiers for Financial Distress Prediction Olmeda,I.;Fernandez,E.
  87. Complex Systems v.4 A genetic learning algorithm for the analysis of complex data Packard,N.
  88. Genetic Algorithms for Pattern Recognition Pal,S.K.;Wang,P.P.
  89. The Accounting Review The Merger Bankruptcy Alternative Pastena,V.;Ruland,W.
  90. Jounal of Finance v.34 no.1 Establishing On-site Bank Examination Priorities: An Early-Warning System using Accounting and Market Information Pettway,R.H.;Sinkey,J.F.
  91. Journal of Finance The Stability of Financial Ratio Pattern in Industrial Organizations Pinches,G.E.;Mingo,K.A.;Caruthers,J.K.
  92. Financial Analysts Journal Firm Mortality: Using Market Indicators to Predict Survial Queen,M.;Roll,R.
  93. The $20^ {th}$ Hawaii International Conference on System Science A Neural Network Application for Bankruptcy Prediction Raghupathi,W.;Schkade,L.L.;Raju.B.S.
  94. Neural networks in the captial markets Refenes,A.P.
  95. Psychological Review v.65 The Perceptron: A probabilistic model for information storage and organization in the brain Rosenblatt,F.
  96. Proceedings of the 3rd International Conference on Artificial Intelligence Applications on Wall Street Experiments with optimal stock screens Rutan,E.
  97. Decision Sciences v.23 Neural networks: A new tool for predicting thrift failures Salchenberger,L.;Cinar,E.;Lash,N.
  98. Proceedings of Second Asia Pacific Decision Sciences Institute (DSI) Conference Using Genetic Algorithm to Support Case-Based Reasoning: Application to Corporate Bond Rating Integration Shin,K.S.;Han,I.
  99. Institute for Operations Research and the Management Sciences A corporate failure prediction model using temporal pattern recognition Shin,K.S.;Hong,S.H.;Lee,S.E.
  100. Bereau of Business Research of University of Illinois Change in the Financial Structure of Unsuccessful Industrial Corporations Smith,R.F.;Winakor,A.H.
  101. Proceedings of $3^ {rd}$ Int'I Conference of Genetic Algorithms Uniform crossover in genetic algorithms Syswerda,G.;Schaffer,J.D.(Eds.)
  102. Journal of Banking and Finance Corporate Bankruptcy Prediction in Japan Takahashi,K.;Kurodawa,Y.;Watase,K.
  103. Management Science v.38 no.7 Managerial applications of neural networks: the case of bank failure predictions Tam.K.;Kiang,M.
  104. Intelligent Systems for Finance and Business Credit evaluation using a genetic algorithm Walker,R.;Haasdijk,E.;Gerrets,M.;S.;Treleaven,P.(Eds.)
  105. Decision Support Systems v.11 no.5 Bankruptcy prediction using neural networks Wilson,R.;Sharda,R.
  106. Journal of Business Research v.44 Probabilistic Neural Networks in Bandruptcy Prediction Yang,A.R.;Platt,M.B.;Platt,H.D.
  107. Journal of Banking and Finance Assessing the vulnerability to failure of Amefican Industrial Firms: A Logistic Analysis Zavgren,C.
  108. Journal of Accounting Research v.22 Methodological Issues Related to the Estimation of Financial Distress Prediction Models Zmijewski,M.E.