Review of Quantitative Finance and Accounting, 17: 351–360, 2001
2001 Kluwer Academic Publishers. Manufactured in The Netherlands.
Fuzzy Numbers in the Credit Rating
of Enterprise Financial Condition
Department of Industrial Engineering, Da Yeh University, Da-Tusen, Chang-Hwa 51505, Taiwan
Department of Industrial Engineering, Yuan Ze University, Chung-Li 320, Taiwan
E. STANLEY LEE
Department of Industrial & Manufacturing Systems Engineering, Kansas State University, Manhattan, KS 66506
Abstract. Most of the parameters used to describe the credit rating are in linguistic terms, which are vague
and difﬁcult to put into precise numerical values. Fuzzy set theory, which was developed to handle this kind of
vagueness, is used to represent and to aggregate the various linguistic data usually used in commercial banks. To
illustrate the approach, numerical examples are solved and compared with existing approaches.
Key words: credit rating, fuzzy sets, linguistic representation
JEL Classiﬁcation: C10, P41
Many factors, which are usually vague, difﬁcult to deﬁne, and even conﬂicting, need to be
considered in determining the credit rating of an enterprise. To make the problem even more
complicated, the relative importance between the different factors must also be decided in
order to obtain the overall rating. Thus, the ﬁnal credit rating obtained is not accurate
and tremendous amount of human judgement is needed in order to use the results. To
overcome this approximate and unreliable nature of credit rating, the so-called rule-of-
thumb is frequently used to screen loan applicants.
Fuzzy set theory was developed to handle this kind of vague and linguistic situation,
and thus is ideally suited for improving the accuracy of credit rating. However, the use
of fuzzy concept in credit rating is fairly new. It appears that Su and Chen (1980) are
the earliest investigators to study this problem using fuzzy sets. Within the last two years,
several researchers have investigated this problem by the use of neural-fuzzy approaches.
Based on the actual data used by the banks in Taiwan, Su and Chen (1980) proposed the
use of fuzzy numbers to represent the various linguistic factors. As will be shown in this
paper, their approaches frequently give unreasonable results. Malhotra and Malhotra (1999)
Address correspondence to: E. Stanley Lee.