摘要
随着社会网络的迅速发展,针对大规模社会网络的可视化已经成为数据挖掘领域中的一项重要的研究课题。传统的布局算法已经无法对大规模的社区网络进行全局管理和展示。因此,该框架基于并行化技术以及分层的思想,实现了大规模社会网络的可视化框架。其贡献主要有:提出了一种基于力导引算法的非重叠社区布局算法(简称NFR);设计了一个基于Spark的并行计算框架;将图数据库(Neo4j)无缝地整合到框架中。最后通过在真实数据集上的测试,验证了该框架的有效性。
With the rapid development of social networks,visualization for large-scale social networks has been an important research topic in the data mining field. The existing layout algorithms failed to manage and demonstrate the large-scale social networks,so the proposed framework realize the visualization framework of large-scale social networks based on parallel techniques and hierarchical ideas. The main contributions include: A non-overlapping community layout algorithm based on force directed layout algorithm( NFR) is proposed. A parallel computing framework based on Spark is designed. A graph database( Neo4j) is seamlessly integrated into the framework. Finally,experiments on various real-world social networks demonstrate the advantage of the framework.
出处
《计算机应用与软件》
2017年第1期73-78,159,共7页
Computer Applications and Software
基金
国家自然科学基金面上项目(71372188)
国家科技支撑计划项目(2013BAH16F01)
关键词
社会网络
力导引布局算法
图数据库
Social networks
Force directed layout algorithm
Graph database