arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.19640cs.SImath.OC
收藏

用于合著超图网络的混合坐标节点链接可视化

Mixed coordinate Node link Visualization for Co_authorship Hypergraph Networks

Mohsen Nafar, Hamed Azami Zenouzagh

AI总结:

本文提出一种混合连续与离散优化的超图节点链接可视化算法,将作者与论文分为不同类别节点并用颜色分类,以提升合著网络布局可读性和分析能力。

摘要:

我们提出了一种算法技术,用于可视化合著网络以及其他以超图(集合系统)建模的网络。由于一篇论文可以由两位以上的研究者共同撰写,研究者之间通过其联合作品产生的交互无法用作者节点之间的直接链接来充分建模。合著网络的超图表示将研究者/作者视为节点,将论文视为超边(作者集合)。我们提出的可视化算法基于一种被广泛研究的方法,该方法将作者和论文都表示为不同类别的节点。我们的方法类似于一些已知的方法,如锚定地图(anchored maps),但引入了一些用于优化顶点定位的特殊技术。该算法同时涉及连续(力导向)优化和离散优化,以确定节点坐标。此外,本工作的一个新颖之处在于使用不同颜色对节点和链接进行分类。这种用法具有明确的目的,有助于观察者从可视化中获取有价值的信息,并提高布局的可读性。该算法经过调优,使观察者能够回答合著网络研究中的特定问题。

英文摘要:

We present an algorithmic technique for visualizing the co-authorship networks and other networks modeled with hypergraphs (set systems). As more than two researchers can co-author a paper, a direct representation of the interaction of researchers through their joint works cannot be adequately modeled with direct links between the author-nodes. A hypergraph representation of a co-authorship network treats researchers/authors as nodes and papers as hyperedges (sets of authors). The visualization algorithm that we propose is based on one of the well-studied approaches representing both authors and papers as nodes of different classes. Our approach resembles some known ones like anchored maps but introduces some special techniques for optimizing the vertex positioning. The algorithm involves both continuous (force-directed) optimization and discrete optimization for determining the node coordinates. Moreover, one of the novelties of this work is classifying nodes and links using different colors. This usage has a meaningful purpose that helps the viewer to obtain valuable information from the visualization and increases the readability of the layout. The algorithm is tuned to enable the viewer to answer questions specific to co-authorship network studies.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑