一种基因本体术语间的语义相似度计算方法

Approach to measure semantic similarity of GO terms

计算基因本体中的术语的语义相似度是基因本体的一个重要应用。基于信息量和基于距离的语义相似度计算方法都只从各自的角度计算术语间语义相似度。提出了基于基因本体中术语所在有向无环图的计算方法。该方法既考虑了术语的祖先对其的信息量的影响,又考虑了术语所在的位置以及术语间的语义联系类型。实验结果表明该方法有较高的准确度。

Calculating the semantic similarity of GO terms is an important application of GO.The information content based and the distance based method only calculate the semantic similarity of terms respectively.This paper proposes a method of directed acyclic graphs of terms to measure the semantic similarity of GO terms.The method not only considers the contribution of ancestor related terms, but also considers the location of term and the semantic relation of terms.Experimental results show that method has higher degree of accuracy.