Abstract

The recent developments in the field of biomedicine have made large volumesof biomedical literature available to the medical practitioners. Due to thelarge size and lack of efficient searching strategies, medical practitionersstruggle to obtain necessary information available in the biomedicalliterature. Moreover, the most sophisticated search engines of age are notintelligent enough to interpret the clinicians' questions. These facts reflectthe urgent need of an information retrieval system that accepts the queriesfrom medical practitioners' in natural language and returns the answers quicklyand efficiently. In this paper, we present an implementation of a machineintelligence based CLINIcal Question Answering system (CLINIQA) to answermedical practitioner's questions. The system was rigorously evaluated ondifferent text mining algorithms and the best components for the system wereselected. The system makes use of Unified Medical Language System for semanticanalysis of both questions and medical documents. In addition, the systememploys supervised machine learning algorithms for classification of thedocuments, identifying the focus of the question and answer selection.Effective domain-specific heuristics are designed for answer ranking. Theperformance evaluation on hundred clinical questions shows the effectiveness ofour approach.

 

Introduction (beta)

None

 

Conclusion (beta)

None