1 AIT Asian Institute of Technology

Similarity measures for resource discovery : an analytical study

AuthorGeorge, K. G.
Call NumberAIT Thesis no.CS-98-19
Subject(s)Information retrieval

NoteA thesis submitted in pa1tial fulfillment of the requirements for the degree of Master of Engineering, School of Advanced Technologies
PublisherAsian Institute of Technology
AbstractThe Intemets size necessitates scalability of algorithms used to locate and track resources. The administrative decentralization of the Internet means that resource information may originate from different sources and formats. The purpose of this work is of threefold; by enumerating the various measures in an 01thogonal framework we make it straight forward for other researchers to describe and discuss similarity measures; by experimenting with a wide range of measures, we hope to observe the features that yield good behavior in a variety of retrieval environment. ; and by describing our results so far, to gather feedback on the issues we have uncovered. The research explored the various facets of similarity measure proposing an eightway orthogonal decomposition into factors that in one form or another appear consistently in most of the literature. This decomposition allows similarity measures to be specified in points of eight-space and so permits the space of similarity measures to be explored in a systematic and coherent manner. Our study is limited to sho1t queries and a subset of the total similarity formulations is chosen for the study. The result shows that Jaccards formulation works well for sho1t queries when factors like document and query term weight, relative term frequencies and document length are taken into consideration.
Year1998
TypeThesis
SchoolSchool of Advanced Technologies (SAT)
DepartmentDepartment of Information and Communications Technologies (DICT)
Academic Program/FoSComputer Science (CS)
Chairperson(s)Sadananda, R.;
Examination Committee(s)Devadason, Francis J. ;Yulu, Qi;
Scholarship Donor(s)Royal Thai Government ;
DegreeThesis (M.Eng.) - Asian Institute of Technology, 1998


Usage Metrics
View Detail0
Read PDF0
Download PDF0