Learning linear classifiers with ternary weights from Metagenomic Data - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2012

Learning linear classifiers with ternary weights from Metagenomic Data

Abstract

Motivated by recent researches in metagenomic classification tasks, this paper investigates the problem of finding interpretable concepts from training data in which the number of features is larger than the number of samples. In this setting, the classification problem is modeled as a combinatorial optimization problem, in which the aim of the learner is to find a {−1, 0, +1}-weighted linear threshold function that minimizes the hinge loss induced from the training data. Two methods are analyzed and experimented: convex optimization that applies randomized rounding to the optimum in the convex hull of the concept class, and supermodular minimization that performs a local search in the concept class, with a guarantee on the approximate solution for the subclass of {0, 1}-weighted threshold functions.
Fichier principal
Vignette du fichier
cap2012_submission_24.pdf (321.75 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-00745449 , version 1 (25-10-2012)

Identifiers

  • HAL Id : hal-00745449 , version 1

Cite

Yann Chevaleyre, Frédéric Koriche, Jean-Daniel Zucker. Learning linear classifiers with ternary weights from Metagenomic Data. Conférence Francophone sur l'Apprentissage Automatique - CAp 2012, Laurent Bougrain, May 2012, Nancy, France. 16 p. ⟨hal-00745449⟩
392 View
135 Download

Share

Gmail Facebook Twitter LinkedIn More