基于概念语义相似度计算模型的信息检索研究

摘要

Taking e-commerce domain ontology as the basis,in the paper we aim at proposing an improved computation model which is based on concept semantic similarity.The model integrates the two aspects of distance-based and content-based,can quantify the semantic similarity between the concept nodes of ontology more comprehensively and accurately.Accordingly,it extends the concept of query keywords set and calculates the similarity of the query and the resulting document,and finally generates the retrieval algorithm as presented in this paper.In experiment we compare it with the Lucene retrieval algorithm,and evaluate the performance of the retrieval algorithm in the paper through choosing key words of focus concept in terms of three indicators:precision,recall and response speed.Experiment proves that the proposed retrieval algorithm in this paper has greater improvement in retrieval performance compared with the information retrieval method based on Lucene.%以电子商务领域本体为基础,旨在提出一种改进的基于概念语义相似度计算模型,该模型结合基于距离和基于内容两个方面,能够更为全面精确地量化本体中概念结点之间的语义相似度.据此,进行查询关键词集概念扩展和查询与结果文档的相似度计算,最终形成检索算法.实验对比于Lucene检索算法,通过选取热点概念关键词从准确率、召回率、响应速度3个指标来评估检索算法的性能.实验证明,提出的检索算法与基于Lucene的信息检索方法相比,检索性能有较大提高.