Title: Automatic annotation generation of medical documents for effective medical information retrieval
Authors: P. Gayathri; N. Jaisankar
Addresses: School of Computing Science and Engineering, VIT University, Vellore-632014, Tamil Nadu, India ' School of Computing Science and Engineering, VIT University, Vellore-632014, Tamil Nadu, India
Abstract: Medical information systems suffer in providing accurate results due to the overwhelmed data. The annotation generation promotes efficient retrieval of medical documents and related concepts. Hence, this paper proposes a novel framework called annotation-based context-aware indexing (ACI) for effective medical information retrieval. In order to manage the diverse nature of the medical terms and user query, informative keywords generated from the medical documents are enriched using Wikipedia and medical ontology. Aggregate the final set of enriched keywords obtained from both Wikipedia and medical ontology to form the annotated keywords. The context-aware indexing using a Bernoulli model improves the performance of information retrieval. It eliminates the irrelevant keywords using the associated value and high-associated keywords are used for indexing medical documents. The proposed ACI achieves better performance than medical ontology based document indexing.
Keywords: medical documents; medical information retrieval; Wikipedia; medical ontology; document indexing; annotation generation; context-aware indexing; medical terms; user queries; keywords; Bernoulli modelling.
International Journal of Reasoning-based Intelligent Systems, 2015 Vol.7 No.3/4, pp.305 - 314
Received: 06 Dec 2014
Accepted: 20 May 2015
Published online: 09 Nov 2015 *