一种面向加权双向图的聚类发掘方法

加权双向图是一种表达具有连接关系的科学和工程问题中的信息的比较准确的方式,加权双向图上聚类发掘的研究具有重要意义.本文提出一种面向加权双向图的聚类发掘方法,它通过定义双向边的调整权和节点的调整度,充分利用加权双向图上边的权值信息和方向信息,比较准确地描述了节点对之间的结构相似性,从而比较好地实现了加权双向图上的聚类发掘.对比实验表明本文方法的聚类发掘结果在聚类质量评价指标上具有更好的表现.

An Approach to Detecting Clusters from Weighted Bidirected Graphs

Weighted bidirected graph is an accurate representation for the problems in science and engineering that have linking relationships.Cluster detection on weighted bidirected graphs is an important research topic.This paper proposes an approach to detecting clusters from weighted bidirected graphs.Based on the definitions of adjusted weight and adjusted degree,this approach measures the structural similarity between nodes accurately,making full use of the weights and direction information of bidirected edges.Therefore,this approach detects clusters from weighted bidirected graphs well.The results of the comparative experiments show that the clusters detected by this approach have better quality than those detected by the comparative approach.