基于多子网复合复杂网络模型的多关系社交网络重要节点发现算法

发布时间：2018-06-02 16:21

本文选题：复杂网络 + 信号传播　；参考：《南京大学学报(自然科学)》2017年02期

【摘要】：为提高社交网络中重要节点评估的效率和有效性,根据社交网络中存在多种相互影响的关系的特性,基于复杂网络理论提出了一种适用于多关系社交网络的重要节点发现算法.首先使用多子网复合复杂网络模型建立包含多种关系的复杂网络,然后采用信号传播方法体现网络中多种关系和节点间的相互影响,以及这些影响的传递性,最后利用矩阵迭代方法模拟信号传播过程,定量分析网络中各节点的重要度.该算法综合了社交网络中节点的全局和局部重要性,考虑了各节点重要度的相互影响,在豆瓣网上的实验结果表明,与传统社交网络重要节点发现算法相比,该算法在迭代次数、涵盖率等指标上都有较大改善,有助于提高社交网络中舆情分析、社团结构发现、信息传播等工作的效率和准确性.
[Abstract]:In order to improve the efficiency and effectiveness of the evaluation of important nodes in social networks, based on the complex network theory, an important node discovery algorithm for multi-relational social networks is proposed. Firstly, a complex network with multiple relationships is built by using a multi-subnet complex network model, and then the signal propagation method is used to reflect the interaction between the various relationships and nodes in the network, and the transitivity of these effects. Finally, the matrix iterative method is used to simulate the signal propagation process to quantitatively analyze the importance of each node in the network. The algorithm synthesizes the global and local importance of nodes in social networks, and considers the mutual influence of the importance of each node. The experimental results on Douban show that compared with the traditional social network important node discovery algorithm, The algorithm can improve the number of iterations, coverage rate and so on, which is helpful to improve the efficiency and accuracy of public opinion analysis, community structure discovery and information dissemination in social networks.
【作者单位】：青岛大学数据科学与软件工程学院;
【基金】：教育部人文社会科学研究青年项目(15YJC86000) 青岛市哲学社会科学规划项目(QDSKL150437)
【分类号】：O157.5

【参考文献】