The Ratio-Cut Polytope and K-Means Clustering
The Ratio-Cut Polytope and K-Means Clustering
复制标题
比率切割多面体和 K 均值聚类
DOI:
10.1137/20m1348601
复制
发表时间:
2022
影响因子:
3.1
通讯作者:
Khajavirad, Aida
中科院分区:
文献类型:
--
作者:
De Rosa, Antonio;Khajavirad, Aida
We introduce the ratio-cut polytope defined as the convex hull of ratio-cut vectors corresponding to all partitions ofpoints ininto at mostclusters. This polytope is closely related to the convex hull of the feasible region of a number of clustering problems such as K-means clustering and spectral clustering. We study the facial structure of the ratio-cut polytope and derive several types of facet-defining inequalities. We then consider the problem of K-means clustering and introduce a novel linear programming (LP) relaxation for it. Subsequently, we focus on the case of two clusters and derive sufficient condition under which the proposed LP relaxation recovers the underlying clusters exactly. Namely, we consider the stochastic ball model, a popular generative model for K-means clustering, and we show that if the separation distance between cluster centers satisfies, then the LP relaxation recovers the planted clusters with high probability. This is a major improvement over the only existing recovery guarantee for an LP relaxation of K-means clustering stating that recovery is possible with high probability if and only if. Our numerical experiments indicate that the proposed LP relaxation significantly outperforms a popular semidefinite programming relaxation in recovering the planted clusters.
登录
查看更多内容
DOI:
--
发表时间:
2019
期刊:
Soft Computing - A Fusion of Foundations, Methodologies and Applications
影响因子:
--
作者:
Francesca Marzi;F. Rossi;S. Smriglio
通讯作者:
S. Smriglio
DOI:
--
发表时间:
1991
期刊:
影响因子:
--
作者:
N. Joshi;R. Dewar
通讯作者:
R. Dewar
影响因子:
2.5
作者:
Hajek, Bruce;Wu, Yihong;Xu, Jiaming
通讯作者:
Xu, Jiaming
DOI:
10.1287/moor.2022.1282
发表时间:
2020
期刊:
Math. Oper. Res.
影响因子:
--
作者:
Alberto Del Pia;Aida Khajavirad;Dmitriy Kunisky
通讯作者:
Dmitriy Kunisky
DOI:
--
发表时间:
2004
期刊:
影响因子:
--
作者:
Jiming Peng;Yu Xia
通讯作者:
Yu Xia