A Study of Credit Rating Models for Small and Medium Enterprises

• Main Authors: Chia-Jung Lu, 呂佳容
• Other Authors: Min-Chun Yu
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/54910942562720357254

碩士 === 國立高雄應用科技大學 === 企業管理系 === 97 === For banking industry, business loan accounts for a large portion of business revenue. However, business profit has been decreasing due to severe price competition. In order to attract more customers, many banks have been either carelessly or knowingly unsuccessful to verify borrowers’ capability of paying back loan on time. As a result, bad debts are piling up and loss incurred. Therefore, the quality of credit rating techniques is vital to effectively grade potential borrowers. Past studies have primarily focused on credit rating methods for large listed companies where rather transparent and accurate financial information has been released regularly. However, for small and medium enterprises, financial information is always hard to collect and with little credibility. This study aims at investigating credit rating models suitable for small and medium businesses. Non-financial information such as human capital for company owners is used to accompany several financial ratios for credit rating. Artificial Neural Networks (ANNs) and Support Vector Machines (SVM) are hence utilized to establish credit rating models using the above information. The performances of the artificial intelligence based techniques are compared with traditional Discriminant analysis.