摘要
由于单一的节点评估方法存在不足,在融合复杂网络的局部特征以及全局特征前提下,提出了一种基于重叠盒覆盖算法的节点重要度评估方法。该方法利用重叠中心性对网络中的节点进行重要度排序,并且与其他不同中心性方法在复杂网络数据集中的节点排序方法进行比较;利用susceptible-infected(SI)模型模拟不同中心性方法前10个节点的传播能力,在此基础上以肯德尔系数进行比较,肯德尔系数越大表明相关性越高。实验结果表明,与其他中心性方法相比,重叠中心性得到的初始节点集合的累积平均感染能力高于其他中心性方法,并且与SI模型具有较高的相关性,该方法对于节点重要度评估是有效并且可行的。
Due to the shortcomings of a single node evaluation method,under the premise of fusing the local and global features of complex networks,this paper proposed a node importance evaluation method based on overlapping box covering algorithm.This method used overlapping centrality to rank the importance of nodes in the network,and compared it with other different centrality methods in complex network data sets.In addition,it used the susceptible-infected(SI)model to simulate the propagation capabilities of the first 10 nodes of different centrality methods,and compared them with the Kendall coefficient on this basis.The larger the Kendall coefficient,the higher the correlation.The experimental results show that,compared with other centrality methods,the cumulative average infectivity ability of the initial node set obtained by overlapping centrality is higher than other centrality methods,and has a higher correlation with the SI model.This method is effective and feasible for the evaluation of node importance.
作者
游倩婧
郑巍
刘方利
You Qianjing;Zheng Wei;Liu Fangli(School of Software,Nanchang Hangkong University,Nanchang 330063,China)
出处
《计算机应用研究》
CSCD
北大核心
2021年第11期3354-3358,共5页
Application Research of Computers
基金
国家自然科学基金资助项目(61867004)
江西省教育厅自然科学基金一般项目(GJJ180523)。
