企業危機預警模型在台灣的應用與比較

• Main Author: 江欣怡
• Other Authors: 張大成
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/24881768976673766836

碩士 === 東吳大學 === 國際貿易學系 === 92 === This paper not only consider the factoers of “base of Accounting” and “base of Market” but also takes macroeconomic factors into account totally including 22 factors and then we use MDA, Probit, Logit and Neural networks to build bankruptcy prediction model. The models that consider the factors, “base of Accounting” and “base of Market,” are called α-models and those with macroeconomic factors are named as β-models. After empirical steps, we find thatβ-models are generally better thanα-models and Logit model perform better than others within in-sample data. Therefore, the models that consider macroeconomic factors really improve the predictive ability of bankruptcy prediction models.