arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1808.04553math.CO
收藏

基于加权团展开的非一致超图的谱界

Spectral bounds for non-uniform hypergraphs using weighted clique expansion

Ashwin Guha, Ambedkar Dukkipati

摘要:

超图是理解大型数据集中许多隐藏模式的宝贵工具。在超图的许多表示方法中,一种有用的表示方法是加权团展开。在本文中,我们考虑这种表示的非一致超图。我们将一致超图的谱结果推广到非一致超图,并证明了它们以一种自然的方式扩展。我们提供了一个边界上的最大特征值相对于平均度的邻居的一个顶点在图中。我们还证明了一个关于顶点集边界的最大和次小特征值不等式,并利用它得到了超图的一些连通性参数的界。

英文摘要:

Hypergraphs are an invaluable tool to understand many hidden patterns in large data sets. Among many ways to represent hypergraph, one useful representation is that of weighted clique expansion. In this paper, we consider this representation for non-uniform hypergraphs. We generalize the spectral results for uniform hypergraphs to non-uniform hypergraphs and show that they extend in a natural way. We provide a bound on the largest eigenvalue with respect to the average degree of neighbours of a vertex in a graph. We also prove an inequality on the boundary of a vertex set in terms of the largest and second smallest eigenvalue and use it to obtain bounds on some connectivity parameters of the hypergraph.

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。

↑