arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1704.02303cs.DS
收藏

网格中的稳定匹配和聚类算法

Algorithms for Stable Matching and Clustering in a Grid

David Eppstein, Michael T. Goodrich, Nil Mamano

摘要:

我们研究了最初由Hoffman、Holroyd和Peres在连续环境中提出的几何稳定婚姻问题的离散版本,其中平面上的点与集群中心稳定匹配,并根据其距离进行优先级排序,以便每个集群中心被分配一组面积相等的点。我们发现,对于一个离散化的问题,$n\times n$网格的像素与$k$中心,这个问题可以解决的时间$O(n^2 \log^5 n)$,我们实验两个速度较慢,但更实用的算法和混合方法,从这些算法之一切换到其他获得更高的效率比单独的算法。我们还展示了如何结合几何稳定匹配的$k$ -手段聚类算法,从而提供一个几何政治选区算法,认为距离在经济方面,我们的实验与加权版本的稳定$k$ -手段,以提高所产生的集群的连接。

英文摘要:

We study a discrete version of a geometric stable marriage problem originally proposed in a continuous setting by Hoffman, Holroyd, and Peres, in which points in the plane are stably matched to cluster centers, as prioritized by their distances, so that each cluster center is apportioned a set of points of equal area. We show that, for a discretization of the problem to an $n\times n$ grid of pixels with $k$ centers, the problem can be solved in time $O(n^2 \log^5 n)$, and we experiment with two slower but more practical algorithms and a hybrid method that switches from one of these algorithms to the other to gain greater efficiency than either algorithm alone. We also show how to combine geometric stable matchings with a $k$-means clustering algorithm, so as to provide a geometric political-districting algorithm that views distance in economic terms, and we experiment with weighted versions of stable $k$-means in order to improve the connectivity of the resulting clusters.

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。

↑