一种融合本体与粗糙集的文档相似度计算方法

An Approach to Compute Document Vector Similarity Based on Rough Set and Ontology

针对向量空间模型VSM中,在计算文档相似度时要求文档标引词必须相互独立这一缺陷,提出融合本体与粗糙集的文档相似度计算方法。在该方法中,不仅可以利用本体对概念关系的准确揭示,计算文档之间的概念相似度,还可以结合粗糙集对相关概念实例的属性重要性进行度量,从而计算属性相似度,克服了传统方法需要依赖人的先验知识这一缺陷,最后综合形成文档语义向量相似度,并通过实验分析证明该方法可以在很大程度上提高文档相似度计算的准确性。

According to the VSM,in the calculation of the document similarity for document indexing words must be independent of the defect,the paper puts forward integration ontology and rough set document similarity calculation method.In this method,not only can use ontology to the conceptual relations accurate reveals that calculation between document concept similarity,but also can combine of rough set,related concept example of attribute importance measure,thus calculation attributes similarity,overcome traditional methods depend on people＇s prior knowledge this defect,and finally the comprehensive form document semantic vector similarity,and through the experimental analysis shows that this method can greatly improve the accuracy of the document similarity calculation.