作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2021, Vol. 47 ›› Issue (8): 109-115,123. doi: 10.19678/j.issn.1000-3428.0059141

• 先进计算与数据处理 • 上一篇    下一篇

注意力流网络中节点影响力的层级性研究

李勇1, 董思秀1, 张强1, 程方颀2, 王常青3   

  1. 1. 西北师范大学 计算机科学与工程学院, 兰州 730070;
    2. 北京航空航天大学 计算机学院, 北京 100190;
    3. 中国互联网络信息中心 互联网基础技术开放实验室, 北京 100190
  • 收稿日期:2020-08-03 修回日期:2020-09-25 发布日期:2020-09-29
  • 作者简介:李勇(1979-),男,副教授、博士,主研方向为大数据、社会计算;董思秀,硕士研究生;张强,副教授、博士;程方颀,学士;王常青,高级工程师、博士。
  • 基金资助:
    国家自然科学基金“基于社会感知计算的公众环境感知与时空行为研究”(71764025);全国高等院校计算机基础教育教学研究项目(2020-AFCEC-355);甘肃省高等学校科研项目(2018A-001);西北师范大学青年教师科研能力提升计划项目(NWNU-LKQN-17-9)。

Research on the Hierarchy of Node Influence in Attention Flow Network

LI Yong1, DONG Sixiu1, ZHANG Qiang1, CHENG Fangqi2, WANG Changqing3   

  1. 1. College of Computer Science and Engineering, Northwest Normal University, Lanzhou 730070, China;
    2. School of Computer Science and Engineering, Beihang University, Beijing 100190, China;
    3. Domain Named System Laboratory, China Internet Network Information Center, Beijing 100190, China
  • Received:2020-08-03 Revised:2020-09-25 Published:2020-09-29

摘要: 复杂网络中节点影响力的层级性在网络结构与控制研究中至关重要。针对有向加权网络中节点影响力的层级性问题,基于海量在线用户行为数据,构建有向加权集体注意力流网络。通过定义节点的层级位置时间和位置约束指标,并结合节点的拓扑位置和时间序列,提出一种用于有向加权网络的节点影响力度量及排序算法。实验结果表明,该算法能有效区分网络层级结构,准确识别出最具影响力的节点,对于节点影响力评估与复杂网络可控性研究具有一定的借鉴意义和参考价值。

关键词: 注意力流网络, 拓扑位置, 时间序列, 节点影响力, K-Shell算法

Abstract: The hierarchy of node influence in complex networks is crucial for the research on network structure and controllability. As for the problem about the hierarchy of node influence in directed and weighted networks, a directed and weighted collective attention flow network is constructed based on massive data of online user behavior. By defining the Hierarchical Position Time(HPT) and position constraint of nodes, an algorithm for measuring and ranking the node influence in directed weighted networks is proposed. The algorithm also considers the topological positions and time series of nodes. Experimental results show that this algorithm can distinguish the hierarchical structure of network, identifying the most influential nodes accurately. It displays certain reference significance and reference value in evaluating the influence of nodes and the controllability of complex networks.

Key words: attention flow network, topological position, time series, node influence, K-Shell algorithm

中图分类号: