Experimental Evaluation of Query Reformulation Techniques in the Context of Medical Information Retrieval

Authors: 
Mohammed Maree
Isra Noor
Khalid Rabayah
ISSN: 
1992-8645
Journal Name: 
Journal of Theoretical and Applied Information Technology
Volume: 
96
Issue: 
24
Pages From: 
8434
To: 
8444
Date: 
Monday, December 31, 2018
Keywords: 
Query Reformulation, Medical Queries, CLEF eHealth Dataset, Medical Knowledge Bases, Precision/Recall Indicators
Abstract: 
With the proliferation of online medical information, a majority of laypeople (ordinary people with little medical background knowledge) and medical specialists now find the Web an indispensable tool for searching for medical information in various domains of interest. However, when using existing medical search engines, the precision of the retrieved results is governed by two main factors. First, users (be they 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 techniques and medical knowledge resources that are exploited by such search engines. Several systems and approaches have been proposed to address problems associated with each of these factors independently; however, little attention has been paid to cooperatively address problems of both factors. In this article, we aim to investigate the impact of exploiting medical knowledge resources and information retrieval techniques on 1) reformulating medical queries through enriching them with semantically-related medical terms and 2) indexing documents in the medical domain to improve the matching process between the reformulated queries and their corresponding medical documents. A prototype of the proposed system has been instantiated and experimentally validated using CLEF2014 eHealth Dataset and state-of-the-art effectiveness indicators. The produced results by our system demonstrate that the quality of the returned results has improved compared to other similar medical information retrieval systems.