An introduction toMajorization‐Minimizationalgorithms for machine learning and statistical estimation
- 9 February 2017
- journal article
- review article
- Published by Wiley in WIREs Data Mining and Knowledge Discovery
- Vol. 7 (2)
Abstract
MM(majorization–minimization) algorithms are an increasingly popular tool for solving optimization problems in machine learning and statistical estimation. This article introduces theMMalgorithm framework in general and via three commonly considered example applications: Gaussian mixture regressions, multinomial logistic regressions, and support vector machines. Specific algorithms for these three examples are derived and Mathematical Programming Series A numerical demonstrations are presented. Theoretical and practical aspects ofMMalgorithm design are discussed.WIREs Data Mining Knowl Discov2017, 7:e1198. doi: 10.1002/widm.1198This article is categorized under: Algorithmic Development > Statistics Technologies > Machine LearningKeywords
Funding Information
- Australian Research Council (ARC)
This publication has 54 references indexed in Scilit:
- MM Optimization Algorithms2016
- Next-Generation Statistical Genetics: Modeling, Penalization, and Optimization in High-Dimensional DataAnnual Review of Statistics and Its Application, 2014
- Optimization2013
- The MM Alternative to EMStatistical Science, 2010
- Graphics Processing Units and High-Dimensional OptimizationStatistical Science, 2010
- The EM Algorithm and Extensions, 2E2008
- A Tutorial on MM AlgorithmsThe American Statistician, 2004
- Optimization Transfer Using Surrogate Objective FunctionsJournal of Computational and Graphical Statistics, 2000
- EM algorithms without missing dataStatistical Methods in Medical Research, 1997
- Support-vector networksMachine Learning, 1995