On the equivalency between decision tree classifiers and the nearest neighbour rule

• J.S. Sánchez ^[1] ; F. Pla ^[1] ; F.J. Ferri ^[2]
  1. [1] Universitat Jaume I

     Universitat Jaume I

     Castellón, España

     
  2. [2] Universitat de València

     Universitat de València

     Valencia, España

     
• Localización: CAEPIA'97: actas / coord. por Asociación Española de Inteligencia Artificial, Vicente J. Botti Navarro , 1997, ISBN 84-8498-765-5, págs. 197-206
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • This paper introduces a method for designing a binary tree structured classifier by using the decision boundaries associated to a reduced set of prototypes. The purpose of this technique consists of finding a classification scheme whose result will be equivalent to that produced by the Nearest Neighbour rule, but with the important advantage of being much faster in terms of classification time. Experiments on real and artificial data sets demonstrate that the design procedure consistently finds decision trees with that equivalency property.