應用資料探勘技術於公司倒閉預測之研究

Predicting Company Bankruptcy Using Data Mining

藍佳樂(Margaret Lungile Dlamini)

指導教授 ： 魏志平

國立清華大學/科技管理學院/科技管理研究所/碩士(2007年)

摘要

Business is very important for the economic growth of every country. The most successful economies have been boosted by growth in the business sector in all industries. There have been mergers and acquisitions and an enormous growth in entrepreneurship in all the successful economies of the developed countries. The developing countries are also following this trend in order to succeed in the global arena. With all these business ventures going on, it is equally important to make sure that the business is carried out in a way that will benefit all stakeholders. But in some instances, the business venture fails because of circumstances beyond their control.

The main objective of bankruptcy prediction is early detection of decline in business activity, whether that decline is caused by factors within the company or outside. Prior studies highlight the most important factors (i.e., financial ratios) to be considered. However, most of them concentrate on the comparison of different statistical or data mining techniques at a single time point by using annual data. This study has addressed the longitudinal analysis of quarterly financial data on bankruptcy prediction. Using 210 companies listed in the Taiwan Stock Exchange (including 105 bankrupt and 105 solvent companies), our empirical results show that the predicting timings closer to the occurrence of actual events will lead to higher effectiveness on bankruptcy prediction. Specifically, prediction made at one or two quarters prior to bankruptcy generally has highest accuracy, precision, and recall rates. In addition, SVM outperforms J48 in the target prediction task. Finally, the inclusion of ratio-change variables has no positive effects on the effectiveness of bankruptcy prediction.

Keywords: Bankruptcy prediction, Data mining, Decision tree induction, Support vector machines (SVM)

關鍵字

bankruptcy prediction ； data mining ； decision tree induction ； support vector machines (SVM)

參考文獻

Abdelwahed, T. and Amir, E.M. (2005). “New Evolutionary Bankruptcy Forecasting Model Based on Genetic Algorithms and Neural Networks,” Proceedings of 17th IEEE International Conference on Tools with Artificial Intelligence (ICTAI 05), pp.241-245.

Almeida, F.C. (1997). “A Structured Approach to Neural Networks in Bankruptcy Prediction,” Proceedings of the 4th Brazilian Symposium on Neural Networks, pp.1-8.

Atiya, A.F. (2001). “Bankruptcy Prediction for Credit Risk Using Neural Networks: A Survey and New Result,” IEEE Transaction on Neural Networks, Vol. 12, pp.929-935.

Back, B., Laitinen, T., and Sere, K. (1996). “Neural Network and Genetic Algorithms for Bankruptcy Prediction,” Expert System with Applications, Vol. 11, No. 4, pp.407-413.

Back, J. and Cho, S. (2003). “Bankruptcy Prediction for Credit Risk Using and Auto-Associative Neural Network in Korean Firms,” Proceedings of 2003 IEEE International Conference Computational Intelligence for Financial Engineering, pp.25-29.