Empirical risk minimization with piecewise-linear classifiers

MaRDI QIDQ1087477zbMATH OpenFDO

Authors E. V. Rozhkov

Publication date 1986

Published in Automation and Remote Control (Search for Journal in Brave)

zbMATH Keywords

mixed integer linear programmingoptimal piecewise-linear classifiers

Mathematics Subject Classification ID

Pattern recognition, speech recognition (68T10) Mixed integer programming (90C11) Applications of mathematical programming (90C90)

Review text

We consider the construction of optimal piecewise-linear classifiers using mixed integer linear programming. A procedure is proposed for constructing piecewise-linear classifiers with simultaneous structure choosing and learning for the case of two classes of patterns. The optimal piecewise-linear classifier for the case of more than two classes is devised. A procedure is developed for the construction of piecewise- linear classifiers in this case.

Recommendations

• Empirical risk minimization and problems of constructing linear classifiers
• Piecewise linear classifiers preserving high local recognition rates.
• Piecewise-linear classifiers and ``generalized moving control: an effective design procedure
• Automated design of multiple-class piecewise linear classifiers
• A novel piecewise linear classifier based on polyhedral conic and max-min separabilities

Cited in

(14)

• Piecewise-linear classifiers and ``generalized moving control: an effective design procedure
• A novel piecewise linear classifier based on polyhedral conic and max-min separabilities
• Classification through incremental max-min separability
• A simple and fast multi-class piecewise linear pattern classifier
• scientific article; zbMATH DE number 3903776 (Why is no real title available?)
• scientific article; zbMATH DE number 1241715 (Why is no real title available?)
• scientific article; zbMATH DE number 1956630 (Why is no real title available?)
• Hinge-minimax learner for the ensemble of hyperplanes
• Empirical risk minimization and problems of constructing linear classifiers
• Piecewise linear classifiers preserving high local recognition rates.
• Optimizing -learning via mixed integer programming
• Learning perceptrons and piecewise linear classifiers sensitive to example dependent costs
• An incremental piecewise linear classifier based on polyhedral conic separation
• Automated design of multiple-class piecewise linear classifiers

This page was built for publication: Empirical risk minimization with piecewise-linear classifiers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1087477)