Sparse weighted voting classifier selection and its linear programming relaxations

DOI10.1016/J.IPL.2012.03.004MaRDI QIDQ436579zbMATH OpenOpenAlexFDO

Authors Noam Goldberg, Jonathan Eckstein

Publication date 25 July 2012

Published in Information Processing Letters (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/j.ipl.2012.03.004

zbMATH Keywords

computational complexityhardness of approximationintegrality gapmachine learningsparsityweighted voting classification

Mathematics Subject Classification ID

Computational difficulty of problems (lower bounds, completeness, difficulty of approximation, etc.) (68Q17) Learning and adaptive systems in artificial intelligence (68T05) Linear programming (90C05) Mixed integer programming (90C11) Applications of mathematical programming (90C90)

Recommendations

• Generalization bounds for voting classifiers based on sparsity and clustering.
• Learning a priori constrained weighted majority votes
• Sparse ensembles using weighted combination methods based on linear programming
• Threshold optimization for weighted voting classifiers
• A weight-adjusted voting algorithm for ensembles of classifiers

Cites work

• 10.1162/153244303322753751
• A decision-theoretic generalization of on-line learning and an application to boosting
• Anti-Hadamard matrices, coin weighing, threshold gates, and indecomposable hypergraphs
• Feature Selection via Mathematical Programming
• Linear programming boosting via column generation
• On the approximability of minimizing nonzero variables or unsatisfied relations in linear systems
• On the hardness of approximating label-cover
• On the Size of Weights for Threshold Gates
• Robust trainability of single neurons
• The Hadamard Maximum Determinant Problem
• The hardness of approximate optima in lattices, codes, and systems of linear equations
• Theory of majority decision elements

Cited in

(5)

• Three algorithms for graph locally harmonious colouring
• Supersparse linear integer models for optimized medical scoring systems
• Learning optimized risk scores
• Sparse ensembles using weighted combination methods based on linear programming
• Combining predictions in pairwise classification: an optimal adaptive voting strategy and its relation to weighted voting

Describes a project that uses

Uses Software

• AdaBoost.MH

This page was built for publication: Sparse weighted voting classifier selection and its linear programming relaxations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q436579)