基于节点相似性的LFM社团发现算法

发布时间：2018-03-12 22:06

  本文选题：复杂网络　切入点：社团发现　出处：《复杂系统与复杂性科学》2017年03期 　论文类型：期刊论文

【摘要】：传统的局部适应度社团发现算法(LFM)在社团结构模糊的网络中精度下降严重。针对此问题,提出LFMJ算法。利用邻居节点信息和改进的杰卡德系数重构网络,使网络结构更为清楚,社团划分结果更为准确。为验证算法,选择了5种算法在LFR网络和真实网络中进行测试,包括LFMJ、LFM和传统的LPA算法以及性能较好的WT和FUA算法。结果表明:在标准LFR网络中,LFMJ精度高于LFM和LPA,与FUA和WT相当;在真实网络和具有重叠结构的LFR网络中,LFMJ精度优于其他4种算法。
[Abstract]:The traditional local fitness community discovery algorithm (LFM) has a serious decline in precision in the network with fuzzy community structure. Aiming at this problem, LFMJ algorithm is proposed to reconstruct the network by using neighbor node information and improved Jekard coefficient to make the network structure clearer. In order to verify the algorithm, five algorithms are selected to test in LFR network and real network. The results show that in the standard LFR network, the accuracy of LFMJ is higher than that of LFM and LPA, which is equivalent to that of FUA and WT. In real network and LFR network with overlapping structure, the accuracy of LFMJ is better than the other four algorithms.
【作者单位】： 信息工程大学理学院理学院;河南中医学院第一附属医院呼吸科;
【基金】：国家自然科学基金(81574100)
【分类号】：O157.5;TP301.6
，

本文编号：1603488