基于灰狼优化算法的社交网络影响力识别算法  

INFLUENCE RECOGNITION ALGORITHM OF ONLINE SOCIAL NETWORKSBASED ON GREY WOLF OPTIMIZATION ALGORITHM

作　　者：李淑霞[1] 杨俊成[1] Li Shuxia;Yang Juncheng(School of Electronic Information Engineering,Henan Polytechnic Institute,Nanyang 473000,Henan,China)

机构地区：[1]河南工业职业技术学院电子信息工程学院,河南南阳473000

出　　处：《计算机应用与软件》2021年第10期297-303,共7页Computer Applications and Software

基　　金：河南省高等学校青年骨干教师培养计划基金项目(2018GGJS230);全国高等院校计算机基础教育研究会课题(2020-AFCEC-196);河南工业职业技术学院青年骨干教师培养计划。

摘　　要：为了提高社交网络中用户影响力识别的准确率,提出一种基于灰狼优化算法的社交网络影响力检测算法。该算法考虑邻居节点和非邻居节点来决定用户间的相似性,并且设计迭代合并的自适应社区检测算法,无需社区数量等先验信息。在用户影响力的识别过程中,采用灰狼优化算法寻找影响力最高的用户,并为灰狼优化算法补充两个变异算子,增加种群的多样性。基于真实数据的实验结果表明,该算法的局部影响力识别准确率和全局影响力识别准确率比现有方法均实现了提高。In order to improve the accuracy of user influence recognition in social networks,we propose an influence detection algorithm of online social networks based on grey wolf optimization algorithm.This algorithm considered both neighbor nodes and other nodes to decide the similarity between two users.And it designed an iterated and merging algorithm to adaptively detect communities,which doesn t need the prior information such as community number.During the procedure of user influence recognition,we used grey wolf optimization algorithm to search the most influential user,and we introduced two mutation operations to grey wolf optimization algorithm to enhance its diversity.The experimental results based on the real data indicate that,the algorithm realizes better recognition accuracies of local influence and global influence compared to existing methods.

关 键 词：灰狼优化算法 变异算子 种群多样性 社交网络 意见领袖 社区检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]