题目：Unsupervised feature selection via Nonnegative Orthogonal Constrained Regularized Minimization
报告人：张立平 教授 （清华大学）
时间：2022年04月28日 下午2：00 开始
地点：腾讯会议 199 761 543
摘要： Feature selection has drawn wide attention in the era of big data since it is one main technique for dimensionality reduction. In this paper, we consider the unsupervised feature selection problem. First, we integrate the feature selection into the nonnegative spectral clustering, and establish the l2,1-regularized minimization with nonnegative orthogonal constraints to model our problem. Second, to solve the nonconvex and nonsmooth optimization model, we propose an inexact augmented Lagrangian multiplier method which hybrids the augmented Lagrangian multiplier method and the proximal alternating minimization method, and show that the sequence generated globally converges to a KKT point of our problem. The numerical experiments on four publicly available datasets illustrate the stability and robustness of our method, and indicate the advantages of our method over the state-of-the-art methods.
报告人信息：张立平，清华大学长聘副教授，博士生导师，研究方向最优化理论算法及应用，在求解互补与变分不等式问题、半无限规划、张量优化等方面取得了一些有意义的结果。已在Mathematics of Computation, SIAM Journal of Matrix Analysis and Applications, SIAM Journal on Optimization, Applied Numerical Mathematics，中国科学等期刊发表高质量论文五十余篇、连续获得多项国家自然科学基金资助。曾获得教育部自然科学奖二等奖和北京市科学技术奖二等奖。
