【摘要】：為提高大型網(wǎng)絡(luò)的社區(qū)發(fā)現(xiàn)精度和效果,解決葉節(jié)點(diǎn)存在的局部極值化問(wèn)題,提出基于局部模塊性增量超點(diǎn)Louvain剪枝技術(shù)的動(dòng)態(tài)社區(qū)發(fā)現(xiàn)方法。首先,對(duì)網(wǎng)絡(luò)社區(qū)進(jìn)行模型定義,并給出社區(qū)發(fā)現(xiàn)的模塊度函數(shù),同時(shí)針對(duì)傳統(tǒng)模塊度函數(shù)存在的葉節(jié)點(diǎn)處置問(wèn)題,對(duì)模塊度函數(shù)進(jìn)行改進(jìn);其次,在進(jìn)行模塊度函數(shù)改進(jìn)基礎(chǔ)上,針對(duì)葉節(jié)點(diǎn)問(wèn)題利用超節(jié)點(diǎn)構(gòu)建Louvain剪枝技術(shù);最后,通過(guò)在社區(qū)發(fā)現(xiàn)算例上實(shí)驗(yàn)對(duì)比顯示,所提算法相對(duì)于對(duì)比算法的模塊度指標(biāo)提升7.2%以上,驗(yàn)證了所提算法有效性。
[Abstract]:In order to improve the accuracy and effect of community discovery in large networks and to solve the problem of local extremum in leaf nodes, a dynamic community discovery method based on local modular incremental super-point Louvain pruning technique is proposed. Firstly, the model of the network community is defined, and the module degree function found by the community is given. At the same time, the module degree function is improved to deal with the leaf node problem existing in the traditional module degree function. Secondly, on the basis of improving the modular degree function, the Louvain pruning technology is constructed by using supernodes to solve the problem of leaf nodes. Finally, the experimental results show that the proposed algorithm is more than 7.2% higher than that of the contrast algorithm, which verifies the effectiveness of the proposed algorithm.
【作者單位】： 四川建筑職業(yè)技術(shù)學(xué)院信息工程系;
【分類號(hào)】：O157.5;TP301.6
，

本文編號(hào)：2409810