一种正方化有序树图布局算法

A Squarified and Ordered Treemap Layout Algorithm

针对现有正方化树图有序性差、有序树图正方化性能差的问题,提出一种正方化有序树图布局算法——SOTLA,首先将多个节点捆绑组合构造大权值节点;然后按平均长宽比最接近1和沿最短边方向填充路径限制的原则对节点进行布局;通过对每一组合节点分治并重复上述过程进行布局,最终完成所有单个节点的布局.将该算法应用于50组具有层次结构和数据值大小悬殊特点的农药残留量检测数据中,得到了较好的可视化效果.将该算法与现有的Squarified,Pivot,Strip和Spiral 4种布局算法就平均长宽比、连续性、可读性、稳定性和稳定的平衡性5项评价指标进行对比,实验结果表明,文中算法在连续性和平均长宽比性能上具有明显优势,其可读性、稳定性和稳定的平衡性指标也表现良好.

Existing squarified treemap layout algorithms are inefficient in sequentiality, and ordered treemap layout algorithms are incapable of squarified ability. In this paper we propose an squarified and ordered treemap layout algorithm, called SOTLA. First, artificial nodes with large values are created by binding nodes. Then, nodes are layouted by letting the average aspect ratio being 1.0 along the shortest edge. Each bundle node is handled in a fashion of divide-and-conquer. The above process is repeated for each node. We test our approach with a dataset of 50 groups of pesticide residues, whose values have large disparity. We compare our approach with conventional Squarified, Pivot, Strip and Spiral layout algorithms in views of average aspect ratio, continuity, readability, and average of distance change and variance of distance changes. Experimental results show that our SOTLA algorithm has significant advantages on continuity and average aspect ratio, and fair performance in terms of readability, average of distance change and variance of distance changes.