论文标题
部分可观测时空混沌系统的无模型预测
Browser-based Hyperbolic Visualization of Graphs
论文作者
论文摘要
双曲线几何形状为数据可视化提供了自然的焦点 +上下文，并已证明是现实世界中复杂网络的基础。但是，当前的双曲网络可视化方法仅限于特殊类型的网络，并且不扩展到大型数据集。考虑到这一点，我们根据反向投影，广义指导算法和双曲线多维缩放（H-MDS）设计，实施和分析了浏览器中网络双曲线可视化的三种方法。与Euclidean MDS的比较表明，H-MDS可为几种类型的网络产生较低失真的嵌入。这三种方法都可以处理节点链接表示形式，并在功能齐全的基于Web的系统中可用。
Hyperbolic geometry offers a natural focus + context for data visualization and has been shown to underlie real-world complex networks. However, current hyperbolic network visualization approaches are limited to special types of networks and do not scale to large datasets. With this in mind, we designed, implemented, and analyzed three methods for hyperbolic visualization of networks in the browser based on inverse projections, generalized force-directed algorithms, and hyperbolic multi-dimensional scaling (H-MDS). A comparison with Euclidean MDS shows that H-MDS produces embeddings with lower distortion for several types of networks. All three methods can handle node-link representations and are available in fully functional web-based systems.
