模糊集合在授信決策上之應用

Applications of Fuzzy Set Theory in Bank Credit Policy

謝海德(Hai-Teh Hsieh)

指導教授 ： 蕭育如

元智大學/工程學院/工業工程與管理學系/碩士(1998年)

摘要

銀行授信行為在商業社會中，一直扮演著重要的角色。而銀行在辦理放款業務時，卻或多或少具有某些程度的風險，也因此評估申貸者之信用是非常重要的。在傳統上，銀行評估借款者主要係依據台北市銀行工會所頒布的授信企業信用評等表，其可分為三大部分：1.財務狀況平評等、2.經營管理、3.產業特性暨展望。第一部分評分之方式乃計算企業之各個財務比率後分別給予1分、2分、……、或5分。這樣的評分方式往往會導致驟變性問題，且不能滿足敏感性。為了克服，本研究乃利用模糊集合論中所定義的梯形模糊數及三角形模糊數，來描述企業財務狀況在各項財務比率中所應得分數之程度，並藉由平均之觀念來做為評分之計算。而經營管理、產業特性暨展望兩部分多為主觀性的語意描述，我們利用模糊理論來處理語言性的描述。更具體來說，我們採用Chen&Hwang所提出的方法，將語言性的描述轉換為模糊數並合成後，再利用Delgado，Vila及Voxman,所提出模糊數之值的計算方式，來表達於此二部分之評分。

關鍵字

信用評等 ； 模糊集合理論

並列摘要

Making loans by commercial banks to firms is an important behavior in financial economy. However, when a bank processes a loan business, it also carries some credit risks. Therefore, assessing the loaner’s credit is critical. Traditionally, a bank evaluates a customer’s credit according to “The credit rating standard for lending to enterprises” published by Committee of the Bank of Taipei City. This standard can be divided into the following three categories: 1.Financial condition of the enterprise, 2.Management, and 3.The particular characteristics of the main product, competition and expectation. In the first category, financial ratios are calculated to evaluate the financial condition of an enterprise and each ratio is represented by five different variables, from 1 to 5. This is due to the problems of anticatastrophism and does not satisfy with the property of sensitivity. Thus, instead of using the five variables, we make the crisp scores into corresponding normal fuzzy numbers, with trapezoidal (or triangular) membership functions. By applying the concept of mean and fuzzy set theory, we proposed a reasonable algorithm to evaluate enterprises financial conditions. No numerical numbers can be obtained for the other two categories and thus only a linguistic representation can be used. It is proven that a fuzzy set theory can handle the linguistic representation, which is generally used in describing the various parameters, effectively. Hence, we convert linguistic terms to fuzzy numbers by Chen&Hwang conversion scale. And then aggregate these fuzzy numbers, and evaluate the value of the fuzzy numbers, which is proposed by Delgado, Vila, and Voxmen, in order to represent the rating of the two categories.

並列關鍵字

bank credit policy ； fuzzy set theory

參考文獻

[15] Adamo, J.M. “Fuzzy Decision Trees,” Fuzzy Sets and Systems, vol. 4, pp. 207-219. 1980.

[16] Baas, S.M., Kwakernaak, H. “Rating and Ranking of Multiple-Aspect Alternatives Using Fuzzy Sets,” Automatic, pp. 47-58. 1977.

[17] Bellman, R.E., Zadeh, L. A. “Decision-maiking in Fuzzy Environment,” Management Sci., No. 4, pp. 141-164. 1970.

[18] Buckley, J.J. “The Fuzzy Mathematics of Finance,” Fuzzy Set and Systems, pp. 257-273. 1986.

[20] Chen, S.H. “Ranking Fuzzy Number with Maximizing Set and Minimizing Set,” Fuzzy Sets and Systems, pp. 113-139. 1985.