基于潜语义分析的概念名称相似度算法

Concept Name Similarity Algorithm Based on Latent Semantic Analysis

概念名称是本体映射中的一个重要要素。针对目前概念名称相似度计算中存在的概念名称多义性问题,提出一种改进的算法。该算法结合概念注释和义项解释,利用潜语义分析,明确概念在Wordnet中对应的义项,在一定程度上提高了相似度计算的准确度。实验表明该方法是切实可行的。

Concept name is important in the process of ontology mapping. This paper proposes a modified concept name similarity algorithm against polysemy of concept name in the process of concept name similarity. Combined with comment and gloss of synset by Latent Semantic Analysis（LSA）, it determines the corresponding synset of concept in the Wordnet. So it improves the accuracy of similarity to a certain extent. Experimental result shows that the algorithm is viable.