Application of Artifical Neural Network to the Financial Distress Predication Models - An Empirical Study of Listed Spinning and Weaving Companies

• Main Authors: Ou, Liang-Feng, 歐良封
• Other Authors: Ouyang, Liang-Yuh
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/29175783491570679939

碩士 === 淡江大學 === 管理科學研究所 === 86 === 　　The purpose of this study isto create a predictive of financial distress which can predict the occurrence of enterprise crisis, and we can help the management of the company to improve the current situation, reduce the possibility of the financial crisis. 　　This paper chooses 26 financial ratios as independent variables and use factor analysis to find representative ratios to be the input variables of prediction models. Then, the neural network''s ability to discriminate between distressed and healthy firm is compared to the traditional statistical model. Result indicate: 1.The closed the year crisis occurred, the financial ratio of the different company are more obvious. 2.the neural network more accurately predicts distressed firms than the discriminate model.