智慧資本對銀行業信用評等之影響

• Main Author: 蔡榮文
• Other Authors: 李天行
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/08815609081303236796

碩士 === 輔仁大學 === 金融研究所 === 91 === Bond ratings play a key role to transfer information in capital market. It could reduce the information asymmetries between lenders and borrowers in the capital markets. In order to meet the above-mentioned objective, credible bond-rating agencies and important information intermediaries, must have good credit analysis technique. In knowledge economy, how to value intangible assets for any business has become a challenging task. Intellectual Capital (IC) represents assets that frequently do not appear in the balance sheet. IC has gained more and more attention since it is the core weapon for many companies. The main purpose of this paper is to explore whether Intellectual Capital will improve the bond rating classification accuracies other than the information provided by financial indices. Two commonly discussed data mining techniques─the backpropagation neural network (BPN) and multivariate adaptive regression splines (MARS) are integrated to build the bond rating classification model using both IC and financial indices. In order to evaluate the bond rating capability of the proposed integrated approach, bond rating task of twenty banks in Taiwan is implemented using the 5-fold cross validation procedure. Empirical results indicate that IC does provide additional information in improving the bond rating accuracies. On the other hand, the proposed approach outperforms traditional BPN and MARS approaches in terms of classification accuracies and misclassification costs.