View : 557 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.author황숙정-
dc.creator황숙정-
dc.date.accessioned2016-08-25T04:08:16Z-
dc.date.available2016-08-25T04:08:16Z-
dc.date.issued2007-
dc.identifier.otherOAK-000000019945-
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/179597-
dc.identifier.urihttp://dcollection.ewha.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000019945-
dc.description.abstractSupport Vector Machine (SVM) 이란 통계학에서 많은 데이터를 처리하기 위해 많이 이용되고 있는 방법 중의 하나이다. 주어진 training set 과 labeled data 그리고 test set 을 이용해서 unlabeled data 를 찾아내는 방법이 transduction 방법으로, 우리는 이 방법에 초점을 맞추도록 하겠다. Semidefinite Method (SDP method) 와 Second Order Cone Programming Relaxation Method (SOCP method) 는 제한조건이 있는 Minimum 또는 Maximum 문제들을 푸는데 효과적인 방법으로 SVM을 푸는데 사용될 수 있다. MATLAB 을 접목해서 SeDuMi (version 1.05) 프로그램을 이용해서 SVM 문제를 풀었고 이로 인해 얻어진 결과는 SOCP method 가 SDP method 보다는 정확도는 약간 떨어지지만, 훨씬 더 효과적이다. 많은 데이터를 단 시간 안에 처리하는 방법으로 SOCP method 가 SDP method보다 효율적이라는 것이 결론이다.;Many statistical methods exist to access the information with data. Support Vector Machine is a well-known method classifying data. Support vector machine (SVM) produces nonlinear boundaries by constructing a linear boundary in a large, transformed version of the feature space. Semidefinite programming is one of the most effective methods to solve nonlinear optimization problems. SVM problems can be solved by semidefinite programming approach. Since SVM problem is a quadratic form and second order cone programming (SOCP) can be applied to enhance the efficiency. Error rate of solving SVM problem with SOCP is lager than with SDP, however, we show that the efficiency can be improved with numerical results.-
dc.description.tableofcontentsAbstract iii Chapter 1 Introduction 1 Chapter 2 Support Vector Machine 3 2.1. Kernel Function 4 2.2. Lagrange Multiplier 5 2.3. Hard Margin 6 2.4. 1-norm Soft Margin - the Box Constraint 8 2.5. 2-norm Soft Margin - Weighting the Diagonal 9 Chapter 3 Semidefinite Relax Programming 12 3.1. Semidefinite Programm 12 3.2. General Optimization 13 3.3. Hard Margin 16 3.4. 1-norm Soft Margin 16 3.5. 2-norm Soft Margin 17 Chapter 4 Second Order Cone Programming 18 Chapter 5 Numerical Experiments 20 5.1. Concluding Remarks 25 References 26 논문 초록 28-
dc.formatapplication/pdf-
dc.format.extent488128 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.titleSupport Vector Machine Via Semidefinite and Second Order Cone Programming Relaxation Methods-
dc.typeMaster's Thesis-
dc.creator.othernameHwang, Sukjung-
dc.format.pageiii, 28 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 수학과-
dc.date.awarded2007. 2-
Appears in Collections:
일반대학원 > 수학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE