欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对规模化网络中局部社团检测存在的对初始节点位置敏感、拓扑信息难以有效利用问题,提出了一种采用影响力节点集扩展的社团检测（IN-LCD）方法。首先定义了节点的局部影响力指标,通过该指标计算并构造了源节点附近的影响力节点子集,然后从影响力节点子集开始,以迭代更新的方式,进行连续的社团扩张,最后通过节点和社团相似性指标计算,完成整个局部社团的获取。IN-LCD方法从有效利用节点局部信息出发,通过最具影响力节点集合进行社团扩展,有效克服了局部社团检测对初始节点位置敏感的问题。在真实和人工网络数据集上的实验表明,IN-LCD方法与已有的最佳局部社团检测方法相比,识别性能提升了5.3%,更能有效应用于局部信息出发的社团检测场景。
英文摘要：
A local community detection algorithm based on influential nodes set（IN-LCD）is proposed to focus the problems that the local community detection in large-scale network is sensitive to the position of source nodes and the topology information is difficult to effectively use.A local influence index for nodes is defined,and a subset of influential nodes near the source node is calculated and constructed with the index.Then,the continuous expansion of the community is realized from the subset,and the whole local community is constructed through the calculation of the similarity index between nodes and community.The method uses the most influential nodes set to expand the community and effectively overcomes the sensitive problem of local community detection to initial node position.Experiments on real and artificial network data sets and a comparison with an existing local community detection method show that the recognition performance of the proposed IN-LCD is improved by 5.3%.
同期刊论文项目
同项目期刊论文
期刊信息
位置：