BoosTexter: A Boosting-based Systemfor Text Categorization

  • Authors:
  • Robert E. Schapire;Yoram Singer

  • Affiliations:
  • AT&T Labs, Shannon Laboratory, 180 Park Avenue, Room A279, Florham Park, NJ 07932-0971, USA. schapire@research.att.com;School of Computer Science & Engineering, The Hebrew University, Jerusalem 91904, Israel. singer@cs.huji.ac.il

  • Venue:
  • Machine Learning - Special issue on information retrieval
  • Year:
  • 2000

Quantified Score

Hi-index 0.02

Visualization

Abstract

This work focuses on algorithms which learn from examplesto perform multiclass text and speech categorization tasks. Ourapproach is based on a new and improved family of boostingalgorithms. We describe in detail an implementation, calledBoosTexter, of the new boosting algorithms for text categorizationtasks. We present results comparing the performance of BoosTexter anda number of other text-categorization algorithms on a variety oftasks. We conclude by describing the application of our system toautomatic call-type identification from unconstrained spoken customerresponses.