Achieving an Appropriate Balance between Precision, Support, and Comprehensibility in the Evolution of Classification Rules

• Autores: Guillermo Leguizamón, Emiliano Carreño Jara
• Localización: Inteligencia artificial: Revista Iberoamericana de Inteligencia Artificial, ISSN-e 1988-3064, ISSN 1137-3601, Vol. 11, Nº. 35, 2007, págs. 23-32
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • This article proposes a method for achieving an appropriate balance between the parameters of support, precision, and complexity during the evolution of classification rules by means of genetic programming. The method includes an adaptive procedure in order to achieve such balance. This work lies within the data mining context, more precisely, it focuses on the extraction of comprehensible knowledge where the approach introduced plays a predominant role. Experimental results demonstrate the advantages of using the proposed method.