透過您的圖書館登入
IP:3.133.141.6
  • 學位論文

類神經集群模式於企業債信評等之應用

NEURAL CLUSTERING FOR CORPORATE

指導教授 : 盧以詮
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


一般運用類神經網路來預測債信等級時,主要預測的目標是Moody’s或是S&P’s所公佈的債信等級。但在國內缺乏公開評等的情況下,若想藉由類神經網路中的監督式學習來進行評等是無法實行的。因此在沒有明確的目標值以供債信評等模式學習與預測的情況下,我們採用了類神經網路中的非監督式學習。在本研究所建立的評等模式中,使用了一個混合自我組織特徵映射(SOFM)與學習向量量化(LVQ)的類神經網路模型,應用於公司債信等級的分類上,此混合SOFM/LVQ系統能增加群集內的同質性與群集間異質性,因而有更佳的分類效果。利用類神經集群模式於企業債信評等非但可以避免評等過程中由專家直接給予權重的程序、並且可以去除產業差異所造成之偏差。另外,本研究以八十四年至八十六年電子產業作一範例,先將電子產業分成六個子產業,再依各子產業以SOFM進行集群分析,而初始集群結果我們以台灣經濟新報因素權重作為給定等級的範例,以提供專家作為賦予等級意義的參考,最後我們以LVQ的方法,將專家調整後的各集群資料建立成模式。

關鍵字

債信評等 類神經 集群模式

並列摘要


When applying neural networks for credit ratings, the targets are the public ratings such as the ones from the Moody’s, or from the Standard & Poor’s. In the absence of public ratings, the application of neural networks by supervised learning or learning by example becomes impractical. We propose a hybrid unsupervised learning model that combines the Self-Organizing Feature Map (SOFM) and the Learning Vector Quantization (LVQ) for the corporation credit ratings classification. The hybrid SOFM/LVQ is implemented as a system to enhance the homogeneity within classes and the heterogeneity between classes. The proposed model avoids the process of giving factors’ weightings by experts. It also reduces the deviation resulting from the differences among industries. We collect the sample data from all listed electronics industry companies in the Taiwan stock market. We use quarterly financial data between 1995 and 1998. Since the electronics companies have their individual characteristics, we divide the electronics industry into six sub-industries. We analyze each of sub-industries using the SOFM. We then provide a set of factors’ weightings of Taiwan Economic Journal as a reference to help experts determine the ratings. Finally, we provide the LVQ using the results learned from the SOFM for future classifications.

並列關鍵字

credit ratings neural networks clustering analysis SOFM LVQ

參考文獻


[1] 李淑娟,「建立我國債信評等制度之構想」,台灣經濟研究月刊,第十八
[8] 林聰明,「台灣地區公司債評等模式之研究」,元智工學院管理研究所,
[17]蘇美芳,「無母數集群模式於企業信用評等之應用」,元智大學資訊社會
[18]Horrigan, J. O. “The Determination if Long-Term Credit
Research, Vol. 4, 1966, pp.44-62.

被引用紀錄


蔡易達(2000)。運用貝氏網路建立台灣地區高風險上市公司之偵測模型〔碩士論文,元智大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0009-0112200611311129
蘇祐萱(2000)。貝氏網路於輔助盈餘預估分析之研究〔碩士論文,元智大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0009-0112200611311909
鄭俊彬(2005)。擴充 EMM 機制以強化 FKMS 於模式功能整合能力〔碩士論文,元智大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0009-0112200611353182

延伸閱讀