In recent years, there has been increasing interest in using text classifiers for retrieving and filtering infomation from web sources. As the numbers of categories in this kind of software applications can be high, Error correcting Output Coding (ECOC) can be a valid approach to perform multi-class classification. This paper explores the use of ECOC for learning text classifiers using two kinds of dichotomizers and compares them to each corresponding monolithic classifier. We propose a simulated annealing approach to calculate the coding matrix using an energy function similar to the electrostatic potential energy of a system of charges, which allows to maximize the average distance between codewords |with low variance. In addition, we use a new criterion for selecting features, a feature (in this specific context) being any term that may occur in a document. This criterion defines a measure of discriminant capability and allows to order terms according to it. Three different measures have been experimented to perform feature ranking/selection, in a comparative setting. Experimental results show that reducing the set of features used to train classifiers does not affect classification performance. Notably, feature selection is not a preprocessing activity valid for all dichotomizers. In fact, features are selected for each dichotomizer that occurs in the matrix coding, typically giving rise to a different subset of features depending on the dichotomizers at hand.

A text classification framework based on optimized error correcting output code

ARMANO, GIULIANO
2015

Abstract

In recent years, there has been increasing interest in using text classifiers for retrieving and filtering infomation from web sources. As the numbers of categories in this kind of software applications can be high, Error correcting Output Coding (ECOC) can be a valid approach to perform multi-class classification. This paper explores the use of ECOC for learning text classifiers using two kinds of dichotomizers and compares them to each corresponding monolithic classifier. We propose a simulated annealing approach to calculate the coding matrix using an energy function similar to the electrostatic potential energy of a system of charges, which allows to maximize the average distance between codewords |with low variance. In addition, we use a new criterion for selecting features, a feature (in this specific context) being any term that may occur in a document. This criterion defines a measure of discriminant capability and allows to order terms according to it. Three different measures have been experimented to perform feature ranking/selection, in a comparative setting. Experimental results show that reducing the set of features used to train classifiers does not affect classification performance. Notably, feature selection is not a preprocessing activity valid for all dichotomizers. In fact, features are selected for each dichotomizer that occurs in the matrix coding, typically giving rise to a different subset of features depending on the dichotomizers at hand.
ECOC classifiers; Feature extraction; Simulated annealing; Computer Science (all)
File in questo prodotto:
File Dimensione Formato  
2015-CEUR-armano.pdf

accesso aperto

Tipologia: versione editoriale
Dimensione 256.31 kB
Formato Adobe PDF
256.31 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11584/197190
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact