网络压缩是一种简明呈现宏观网络结构的重要网络可视化技术,针对现有压缩方法在节点划分中存在不公平的现象,为了快速呈现不同规模的层次结构,提出一种基于非重要节点拆分融合的网络层次压缩算法.首先在定义节点块及其关系的基础上提出针对节点块的拓扑结构重要性度量方法;然后利用资源分配和分流原理,提出基于节点拆分的拓扑层次聚合机制方法;最后通过将网络中每一层的非重要节点块拆分融合到与其相邻的节点块中,实现网络结构的快速压缩.实际网络测试结果表明,该算法能够克服传统方法在节点划分中导致的不公平问题;相比其他方法,其能够展现不同层次的网络结构,并具有良好的层次可视化效果.%As an important network visualization technique, network compression can show the global net-work structure of large scale networks. To solve the unfair problem in partitioning and show the hierarchical structures quickly with different scales, this paper proposes a hierarchical network compression algorithm based on splitting and merging unimportant nodes. The algorithm firstly proposes a method to measure the importance of different node blocks based on the definition of node blocks and connections between them. Then, motivated by the process of re-source allocation, it proposes a hierarchical aggregation mechanism based on node splitting. Finally, it can quickly compress the network by splitting and merging unimportant nodes in each hierarchical structure compression. Empirical study on several real networks shows that the proposed algorithm can solve the unfair problem in partitioning and achieve a good performance on visuali-zation with multiple hierarchical compression structures, compared with other methods.
展开▼
