Asia-pacific Journal of Industrial Management

　　《ASIA-PACIFIC JOURNAL OF INDUSTRIAL MANAGEMENT》中：Abstrtct This paper devehlps a case based rea,soning system to evaluate customers' credit in a smallcompany. The system extracts feature data of customersfrom a management information system databaseand assesses the credit score of a target customer on the basis of similarity to cases already stortxl.

　　本书主要内容：This first issue of ＜Asia Pacfic Journal of Industrial Management＞ is a selected edition of the 8th International Conference on Industral Management （ICIM） proceeding，covering research in industrial production management，advanced production management techniques，business administration man-agement，project management，research and development （R&D），industri-al quantitative analytical methodologies，and theroretical／empirical studies on artificial intelligence in industrial management，providing versatile refer-ences to specialists and experts as well as academic scholoars in industrial management and related areas.

　　Abstrtct This paper devehlps a case based rea,soning system to evaluate customers' credit in a smallcompany. The system extracts feature data of customersfrom a management information system databaseand assesses the credit score of a target customer on the basis of similarity to cases already stortxl. A case is represented as a dataset ctmtaining eight features such asthe type of customers, sales, amount of overdue payment, etc. The procedure for similar cases retrievingand credit scoring program based on similar cases areproposed. The perfornmnce and effectiveness of thesystem is confirmed hy applying it to real credit evaluaring prohlems of the company. The credit scores for sound customers provided hy the system are more than98% in agreement with the judgments of financial managers of the company. The system outperforms, discriminant analysis methods and it has a higher ability to discriminate sound customers from unsound ones.
　　In today's increasingly competitive business environmento successful risk management is very important.Credit risk is most simply defined as the potential that the counterparty will fail to meet its obligationsin accordance with agreed terms. The goal of credit risk management is to avoid the loss encounteredin the counterparty's failures or bankruptcies. Thecredit risk over a given la'riod involves both the prohability of defaulting and the fraction of the value that islikely to be lost in the event of default. Because there are many types of counter parties from individuals tosovereign governments and many different types of obligations from auto loans to derivatives transactionscredit risk takes many forms.
　　There has been significant prior research on creditanalysis and credit evaluation. The mcah'ls and methodologies published so far for credit risk assessmentgenerally fall into two categories:default models and credit soring models. Defauh mtxiels assess the likelihtood of an ohligor's defaulting. Credit scoring models are used to assess the credit qualhy of counterparty.