Semantic Similarity Measures to Disambiguate Terms in Medical Text

Abstract

Computing the semantic similarity accurately between words is an important but challenging task in the semantic web field. However, the semantic similarity measures involve the comprehensiveness of knowledge learning and the sufficient training of words of both high and low frequency. In this study, an approach MedSim is presented for semantic similarity measures to identify synonym terms in medical text with effectiveness and accuracy well-balanced. Experimental results on Chinese medical text demonstrate that our proposed method has robust superiority over competitors for synonym identification.