View : 58 Download: 0

An Application of Support Vector Machines for Corporate Bond Rating Modeling

Title
An Application of Support Vector Machines for Corporate Bond Rating Modeling
Authors
김지경
Issue Date
2005
Department/Major
대학원 경영학과
Keywords
Bond Rating PredictionSupport Vector MachinesBack-propagation Neural Network
Publisher
이화여자대학교 대학원
Degree
Master
Advisors
SHIN, Kyung-Shik
Abstract
This strudy presents an application of Support Vector Machines (SVM) for the prediction of corporate bond rating. For bond rating, traditional statistical methods and artificial intelligence methods have been used. Among them, Back-Propagation Neural Network (BPN) has shown better performance than others. However, BPN needs data as much as possible to develop a model and has many parameters to be controlled. On the other hand, SVM captures geometric characteristics of feature space without deriving weights of networks from the training data, it is capable of extracting the optimal solution with a small number of training data. In case of multiple classification problems, there are classes that have a small number of samples compared with other classes. Thus, we applied SVM to prediction of corporate bond rating, and compared the result with BPN. The results demonstrate that SVM is effective to classify groups that have a small number of data.;본 연구는 통계적 방법론을 기반으로 하는 Support Vector Machine을 활용하여 기업 채권등급 평가를 위한 모형을 구축하였다. 기업의 채권등급 평가 모형 구축을 위해 전통적으로 통계적 기법과 인공지능적 접근 방식이 사용되어 왔다. 이 중 많은 연구에서 우수한 성능이 검증된 Back-propagation neural network과 SVM의 결과를 총 적중률과 그룹별 적중률로 나누어 비교하였으며, 실험 결과 적은 수의 데이터를 갖는 등급의 분류에 있어서 SVM이 좀 더 효과적이라는 사실을 알 수 있었다. 본 연구는 SVM을 기업 채권등급 평가 모형에 적용함으로써 SVM의 활용가능성을 보였다는데 그 의의가 있다.
Fulltext
Show the fulltext
Appears in Collections:
일반대학원 > 경영학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE