INTEGRATING FINANCIAL RATIOS AND CORPORATE GOVERNANCE INDICES TO BUILD THE MODEL OF CREDIT RATING PREDICTION—APPLICATION OF MULTIVARIATE DISCRIMINATE ANALYSIS AND ARTIFICIAL NEURAL NETWORK

• Main Authors: LO, YU-HUI, 羅玉惠
• Other Authors: GOO, YEONG-JIA
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/23148047855524939097

碩士 === 國立臺北大學 === 企業管理學系 === 95 === After the Asian financial crisis in 1997, many well-known enterprises faced with operating crisis and lost investors’ confidence. One of the main reasons is the mechanism of corporate governance not sound enough. Many researches found non-financial information can reflect business crisis better than financial information. The function of credit rating is to evaluate the company’s ability to meet its financial obligations. So it can be seen an indicator of business financial crisis, especially many Tank stocks appear recently in our country. So our research wants to discuss how the non-financial information, corporate governance variables can predict corporate credit rating, and compare difference between two models built by two distinct forecast technologies. The empirical results are as the followings: First our research uses multivariate discriminate analysis to build a predicting model and screen key variables. The empirical result is that the hit ratio of integrating financial ratios and corporate governance indices model is better. Moreover we use Genetic Algorithm extracting final 9 variables with heavy impact on credit ratting result. Besides, more variables belong to corporate governance indices that mean corporate governance is the important information source of business evaluation. Second we compare the forecasting ability of Multivariate Discriminate Analysis model with Artificial Neural Network model. We find the latter model built only by 9 variables but its whole validity (90% hit ratio), internal validity (89.29% hit ratio) and external validity (88.57% hit ratio) all are better than Discriminate Analysis model. In contrast to two models, Artificial Neural Network model have better generality that can provide external stakeholders to apply different sample businesses to forecast risk degree.