基于社区探测的层次自适应并行布局算法

针对大规模网络高效布局和递进式结构分析的需求,提出基于社区发现的多层级力导向布局算法.首先,该算法采用Louvain算法对网络进行多层级社团结构划分,根据划分结果压缩网络并进行骨架布局,确定网络整体架构;然后,采用自适应的力导向变体算法对各个社团内部的原始节点并行布局,细化社区内部网络结构,并引入补偿力减少社区划分带来的网络结构信息缺失;最后,设计了初始布局算法、改良了振颤模型来减少布局所需的迭代次数.实验结果表明,与现有网络布局算法相比,该算法能够更清晰、高效地展示大规模社交网络数据,满足大规模复杂网络可视化的需要.

Hierarchical adaptive parallel layout algorithm based on community probe

To meet the requirement of efficient layout and progressive structure analysis of a large-scale network, this paper proposed a multi-hierarchy force-oriented layout algorithm based on community discovery. First of all, the algorithm employed the Louvain algorithm to divide the network into a multi-level community structure. According to the results of the division, it compressed the network and laid out the framework to determine the overall network architecture. Then, it employed the adaptive force-oriented variant algorithm to arrange the original nodes in parallel within each community, refined the internal network structure of each community, and introduced the compensation force to reduce the lack of network structure information caused by community division. Finally, it designed an initial layout algorithm and improved a tremor model to reduce the number of iterations required for layout. The experimental results show that compared with the existing network layout algorithm, the proposed algorithm displays large-scale social network data more clearly and efficiently, and meets the needs of large-scale social network visualization.