基于随机游走理论的改进LFM算法

发布时间：2018-04-21 02:38

本文选题：复杂网络 + 社团发现　；参考：《计算机工程》2017年11期

【摘要】：传统LFM社团发现算法基于网络局部信息进行社团划分,未充分利用网络中包含的结构信息,导致社团结构模糊的网络中社团划分精度下降严重,同时算法基于局部信息扩张社团,容易形成畸形社团结构。为解决上述问题,提出一种改进的LFM算法,利用随机游走理论衡量节点相似度,使社团结构更清晰,同时寻找赋权网络中的极大子团,以子团为基本单位进行社团扩张,解决畸形社团问题。在人工网络和真实网络上的实验结果表明,与传统LFM算法、标签传播算法等相比,改进的LFM算法具有更高的社团划分精度。
[Abstract]:The traditional LFM community discovery algorithm is based on the local information of the network to divide the community, which fails to make full use of the structure information contained in the network, resulting in a serious decline in the accuracy of community division in the fuzzy network. At the same time, the algorithm is based on local information to expand communities, easy to form a deformed community structure. In order to solve the above problems, an improved LFM algorithm is proposed to measure the similarity of nodes using random walk theory to make the community structure more clear. At the same time, we find the maximal sub-clusters in the weighted network and expand the communities with sub-clusters as the basic unit. Solve the problem of deformities. The experimental results on artificial and real networks show that the improved LFM algorithm has higher community division accuracy than the traditional LFM algorithm and label propagation algorithm.
【作者单位】：信息工程大学理学院;河南中医大学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

【相似文献】