Data filtering for cluster analysis by _0-norm regularization

DOI10.1007/S11590-017-1152-7zbMATH Open1385.90025arXiv1607.08756OpenAlexW3102984106MaRDI QIDQ1686543FDOQ1686543

Andrea Cristofari

Publication date: 15 December 2017

Published in: Optimization Letters (Search for Journal in Brave)

Abstract: A data filtering method for cluster analysis is proposed, based on minimizing a least squares function with a weighted $el{l}_0$-norm penalty. To overcome the discontinuity of the objective function, smooth non-convex functions are employed to approximate the $el{l}_0$-norm. The convergence of the global minimum points of the approximating problems towards global minimum points of the original problem is stated. The proposed method also exploits a suitable technique to choose the penalty parameter. Numerical results on synthetic and real data sets are finally provided, showing how some existing clustering methods can take advantages from the proposed filtering strategy.

Full work available at URL: https://arxiv.org/abs/1607.08756

Recommendations

• Sparse $k$-Means with $\ell_{\infty}/\ell_0$ Penalty for High-Dimensional Data Clustering
• An algorithm for minimizing clustering functions
• Center-based \(l_1\)-clustering method
• A continuous exact \(\ell_0\) penalty (CEL0) for least squares regularized problem
• A DC optimization approach for constrained clustering with \(\ell_1\)-norm

zbMATH Keywords

cluster analysisnonlinear optimizationzero-norm approximation

Mathematics Subject Classification ID

Applications of mathematical programming (90C90) Nonlinear programming (90C30)

Cites Work

• Estimating the number of clusters in a data set via the gap statistic
• Title not available (Why is that?)
• Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties
• Finite mixture models
• Title not available (Why is that?)
• Variational Analysis
• Title not available (Why is that?)
• Likelihood-Based Selection and Sharp Parameter Estimation
• Title not available (Why is that?)
• Title not available (Why is that?)
• A nonmonotone truncated Newton-Krylov method exploiting negative curvature directions, for large scale unconstrained optimization
• Data Clustering: Theory, Algorithms, and Applications
• Title not available (Why is that?)
• Concave programming for minimizing the zero-norm over polyhedral sets
• Handling missing values in support vector machine classifiers
• Title not available (Why is that?)
• Solution path clustering with adaptive concave penalty
• Title not available (Why is that?)

Uses Software

• UCI-ml
• clusterpath

This page was built for publication: Data filtering for cluster analysis by \(\ell _0\)-norm regularization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1686543)