EFFICIENT DIVERSITY AWARE RETRIEVAL SYSTEM FOR HANDLING MEDICAL QUERIES

V.Sudha, E.V.R.M. Kalaimani

Abstract: Clinical research are meagre in healthcare today and most clinical data analytical tools have difficulty in extracting valid information from unstructured data.The hike of Electronic Medical Records (EMRs)causes an eruptive growth of the medical data. The current paper based medical search technologies are tedious to find meaningful patient information in the large medical database. Moreover, the good quality medical search in large collection of EMR is a challenging task especially due to the complex semantic relationships among medical concepts. Hence, to address the semantic issues, asearch result diversification and semantic based IR are proposed. Meta map concept identifier is introducedto map the biomedical query terms to the corresponding MeSH terms. With the help of clinical domain ontology, all the potential semantics from an input query are mined and consumed to model in to variety of query aspects. The extent of VSM is used to constitute the EMR log as vectors. Each group of words consists of multiple concepts and words. The documents are ranked based on its importance. In this project, aNeural Network based Classifier is used to train and test the medical documents and queries topredict the similarity measure between them.

Keywords: Electronic Medical Records, Information Retrieval, Meta Map Concept Identifier, MeSH Ontology, Neutral Network Classifier

DOI: https://doi.org/10.15623/ijret.2016.0503065