Construction of a credit scoring model using data envelopment analysis and support vector machine for evaluating and improving customers’ contribution.

• Main Authors: Lee Chung-Ta, 李忠達
• Other Authors: Tian-Shyug Lee, Chi-Jie Lu
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/23021583531464559740

碩士 === 輔仁大學 === 管理學研究所 === 96 === Clustering and Classification are the most popular techniques in credit scoring. Most of the credit scoring models are two stages combined by clustering and classification. However, most of the hybrid models are lack of revising abilities. To overcome this limitation, a two stages credit scoring model using DEA and SVM is proposed. Data envelopment analysis (DEA) is a clustering model based on the efficiency score－the “contribution score”, the most important advantage of DEA is providing an indeed improvement for decision making unit (DMU). Support vector machine (SVM) is a classification model based on the statistical learning theory and has successfully applied to wide range of application. In this study, the experiment is based on the credit card database of the bank. According to the efficiency score based on the DEA model, the experiment will distinguish the efficient customers from non-efficient ones and create the classified model with SVM. This study will also analyze the non-efficient customers by DEA to give an improvement which can improve or revise the bank policy in order to promote the one’s efficiency. The experiment shows that the hybrid model of this paper proposed can provide an indeed improvement for bank to improve the efficiency of non-efficient customer.