利用互斥策略优化二分网络节点预测  

Optimization of bipartite network node prediction using mutual exclusion strategy

作　　者：范纯龙[1,2] 王洁琼 范东皖 丁国辉 FAN Chun-long;WANG Jie-qiong;FAN Dong-wan;DING Guo-hui(Large-scale Distributed System Laboratory of Liaoning Province,Shenyang Aerospace University,Shenyang 110136,China;School of Computer Science,Shenyang Aerospace University,Shenyang 110136,China)

机构地区：[1]沈阳航空航天大学辽宁省大规模分布式系统实验室,沈阳110136 [2]沈阳航空航天大学计算机学院,沈阳110136

出　　处：《沈阳航空航天大学学报》2019年第5期48-59,共12页Journal of Shenyang Aerospace University

基　　金：国家自然科学基金项目(项目编号:61303016)

摘　　要：目前,网络节点预测的相关研究主要集中在两个方面:一是对网络中“源头节点”的定位;二是对“隐藏节点”的发现,而缺少关于新生节点预测方向的研究。针对该现状,通过采集期刊上面的论文及对应的关键词信息,构建了论文-关键词二分网络数据集,然后将二分网络加权投影为关键词关系网络,并在该网络上利用关键词组合情况预测新论文节点的产生。发现节点间存在一种互相排斥的现象,从另一方面描述网络节点间关系,结合改进的相似性算法,提出新生节点预测算法。利用论文-关键词二分网络实验数据集,对提出的节点预测算法进行验证,并与已有算法作对比,结果显示该节点预测算法对预测新生论文的产生有着更好的效果。At present,the research on network node prediction mainly focuses on two aspects:one is the positioning of the source node;the other is the finding of the hidden node.However,there is a lack of research on the prediction of new nodes.To this end,this paper constructs a bipartite network as the data set by collecting the information of journal papers and keywords.Then,the bipartite network is weighted and projected into a keyword relation network,and the generation of new paper nodes is predicted by using keyword combinations.Moreover,it is found that there is a mutually exclusive phenomenon between nodes,and the relationship can be described from another aspect.By combining an improved similarity algorithm,a new node prediction algorithm is proposed.Based on the constructed experimental data set,the proposed algorithm is verified and compared with the existing algorithms.The results show that the proposed algorithm has better prediction effects.

关 键 词：节点预测 链路预测 二分网络 加权投影 互斥性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]