arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1802.07091math.OCcs.LG
收藏

一种有效的半光滑牛顿凸聚类算法

An Efficient Semismooth Newton Based Algorithm for Convex Clustering

National University of Singapore(新加坡国立大学);Hong Kong Polytechnic University(香港理工大学)

Yancheng Yuan, Defeng Sun, Kim-Chuan Toh

摘要:

聚类可能是无监督学习中最基本的问题,由于其在许多应用中的重要性,无监督学习在机器学习研究中仍然很活跃。像K-means这样的流行方法可能会受到不稳定性的影响,因为它们容易陷入局部最小值。最近,在[7]和[5]中提出了范数和(SON)模型(也称为聚类路径),这是层次聚类模型的凸松弛。虽然提出了ADMM和AMA等数值算法来求解凸聚类模型[2],但众所周知,解决大规模问题非常具有挑战性。针对大规模凸聚类问题,提出了一种基于半光滑牛顿的增广拉格朗日方法。模拟和真实数据的大量数值实验表明,我们的算法是高效和强大的解决大规模问题。此外,数值结果也显示了优越的性能和可扩展性,我们的算法相比,现有的一阶方法。

英文摘要:

Clustering may be the most fundamental problem in unsupervised learning which is still active in machine learning research because its importance in many applications. Popular methods like K-means, may suffer from instability as they are prone to get stuck in its local minima. Recently, the sum-of-norms (SON) model (also known as clustering path), which is a convex relaxation of hierarchical clustering model, has been proposed in [7] and [5] Although numerical algorithms like ADMM and AMA are proposed to solve convex clustering model [2], it is known to be very challenging to solve large-scale problems. In this paper, we propose a semi-smooth Newton based augmented Lagrangian method for large-scale convex clustering problems. Extensive numerical experiments on both simulated and real data demonstrate that our algorithm is highly efficient and robust for solving large-scale problems. Moreover, the numerical results also show the superior performance and scalability of our algorithm compared to existing first-order methods.

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。

↑