中文摘要：

提出一种多变元网络可视化方法MulNetVisBasc,根据节点的多变元属性,使用高级星形坐标法布局网络节点,以边融合及路由技术为基础设计算法,自动有效布局网络边,实现友好的人机交互界面辅助用户进一步对数据进行分析挖掘.实验结果表明,MulNetVisBasc的可视化结果能够在直观揭示数据集多变元分布特性的同时清晰展现其网络关联特性,有助于用户发掘多变元网络数据集中潜在的隐性知识.边布局算法能够有效减少视图中的边交叉数量,且复杂度较低,适用于较大规模数据集,人机交互界面灵活方便.

英文摘要：

This article proposes a multivariate network visualization paradigm, MulNetVisBasc. Advanced Start Coordinates （ASC） are employed to place nodes on the basis of multivariate attributes and to devise an algorithm that that incorporates edge-merging and routing techniques to automatically lay-out edges; furthermore, a user-friendly human-computer interface is developed to assist users in further data analysis and mining. The experimental results suggest that the visualization of MulNetVisBasc not only uncovers the multivariate distributional characteristics of datasets intuitively, but also displays the associations of networks clearly and is helpful in discovering the implicit knowledge hidden behind datasets. The edge layout algorithm reduces the visual clutters caused by edge crossing and is suitable for relatively huge multivariate network datasets in virtue of its low complexity. Finally, the human-computer interface is flexible and convenient.