지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학술대회자료
- 저자정보
- 발행연도
- 2014.5
- 수록면
- 363 - 370 (8page)
이용수
초록· 키워드
오류제보하기
Many studies have focused on the prediction of corporate credit rating using various data mining techniques. The most frequently-used algorithms are SVM and NN. The emergence of novel techniques SVM+ and SVM+MTL enriched data mining techniques and provided one more alternative choice to application. In this paper, we first applied SVM+ and SVM+MTL to corporate credit rating in the context of multi-classification problem. This study tried to solve the multi-class SVM+ and SVM+TML problems by constructing several binary classifiers. To examine the robustness of performance for SVM+MTL algorithm, we utilized 4 different kinds of typical multi-class processing methods in our experiments. Additionally, we compared the results of SVM+MTL with the conventional SVM, NN and MDA methods. The results showed that SVM+MTL outperformed all the other methods, and we proposed SVM+MTL technique as the first choice.
목차
Abstract
Ⅰ. Introduction
Ⅱ. Literature review
Ⅲ. Research framework
Ⅳ. Expriments & results analysis
Ⅴ. Conclusions
References
참고문헌 (0)
참고문헌 신청이 논문의 저자 정보
최근 본 자료
댓글(0)
0
UCI(KEPA) : I410-ECN-0101-2016-003-001367844