Show simple item record

dc.contributor.authorLang, Nguyen
dc.date.accessioned2020-04-24T14:54:57Z
dc.date.available2020-04-24T14:54:57Z
dc.identifier.urihttp://hdl.handle.net/10222/78831
dc.description.abstractIn many real-world applications, a high number of features could result in noisy and redundant information, which could degrade the general performance of classification tasks. Feature selection techniques with the purpose of eliminating such features have been actively studied. In several information-theoretic approaches, such features are conventionally obtained by maximizing relevance to the class while the redundancy among the features used is minimized. This is an NP-hard problem and still remains to be a challenge. This research proposes an alternative feature selection strategy on binary text representation data based on the properties of submodular functions, with the purpose of providing a theoretical lower bound for finding a near optimal solution based on the Maximum Relevance-Minimum Redundancy criterion. In doing so, the proposed method can achieve a 2-approximation by a naive greedy search. Empirical experiments validated and benchmarked against different baseline methods show that the proposed technique is a promising approach on binary data in general.en_US
dc.language.isoen_USen_US
dc.subjectFeature selectionen_US
dc.subjectSubmodularen_US
dc.subjectClassificationen_US
dc.titleA NEAR OPTIMAL SOLUTION FOR MAXIMUM RELEVANCE MINIMUM REDUNDANCY FEATURE SELECTIONen_US
dc.typeThesisen_US
dc.date.defence2020-04-09
dc.contributor.departmentFaculty of Computer Scienceen_US
dc.contributor.degreeMaster of Computer Scienceen_US
dc.contributor.external-examinern/aen_US
dc.contributor.graduate-coordinatorDr. Michael McAllisteren_US
dc.contributor.thesis-readerDr. Evangelos Miliosen_US
dc.contributor.thesis-readerDr. Vlado Keseljen_US
dc.contributor.thesis-supervisorDr. Nur Zincir-Heywooden_US
dc.contributor.ethics-approvalReceiveden_US
dc.contributor.manuscriptsNoen_US
dc.contributor.copyright-releaseNoen_US
 Find Full text

Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record