Please use this identifier to cite or link to this item: http://repository.aaup.edu/jspui/handle/123456789/2858
Title: QUERY REFORMULATION IN THE CONTEXT OF MEDICAL INFORMATION RETRIEVAL رسالة ماجستير
Authors: Noor, Isra' A.$AAUP$Palestinian
Keywords: media system,research step,research questions,problem statement
Issue Date: 2018
Publisher: AAUP
Abstract: In the recent years, search engines on the Web have become indispensable tools that assist users in finding information in almost any domain of interest. These search engines can be classified into two main broad categories; generic (known as horizontal) and domain-specific (known as vertical) search engines. Both types of search engines exploit different algorithms and techniques to first index documents on the Web and accordingly retrieve the most relevant results that satisfy users’ information needs. In this research work, our focus has been on the second category of search engines i.e. domain-specific search, namely in the medical domain. When using medical search engines, the precision of the retrieved results is mainly dependent on two major factors. First, users (laypeople or medical professionals) need to submit vocabularies that best describe their information needs. Second, the quality of the returned results is largely based upon the effectiveness of the employed techniques and medical knowledge resources used by such search engines. Several systems and approaches have been proposed to address problems associated with each of these factors independently; such as using general terms to express laypeople needs instead of using professional medical concepts, and using knowledge resources with limited coverage of medical concepts which lead to a semantic gap between user's queries and their related medical documents. However, little attention has been paid to jointly address problems in both factors. In this research, we have planned to investigate the impact of exploiting medical semantic resources and information retrieval techniques on 1) reformulating the given users’ queries through enriching them with semantically related terms and 2) indexing documents in the medical domain to improve quality of the v retrieved search results. To demonstrate the effectiveness of the proposed system, we have conducted several experiments using the Conference and Labs Evaluation Forum (CLEF) eHealth2014 dataset. Additionally, we have evaluated the effectiveness of the proposed system (MedIR) against three state-of-the-art medical information retrieval systems - that used the same dataset for system evaluation - based on the well-known precision indicator. The conducted experiments using the above dataset show that the proposed approach has achieved a precision of 0.7945, while the precision of the other similar medical information retrieval systems we compared with was (0.7560, 0.7540, and 0.7400). The previous precision values demonstrate that the quality of the retrieved results of our approach has improved by around five to seven percent compared to other similar systems.
Description: Master`s degree in Computer Science
URI: http://repository.aaup.edu/jspui/handle/123456789/2858
Appears in Collections:Master Theses and Ph.D. Dissertations

Files in This Item:
File Description SizeFormat 
اسراء نور.pdf1.34 MBAdobe PDFThumbnail
View/Open


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

Admin Tools