COGITO 2.0una herramienta para obtener un clasificador jerárquico en aprendizaje supervisado

  1. José Riquelme 1
  2. Jesús Aguilar 1
  3. Miguel Toro 1
  1. 1 Universidad de Sevilla
    info

    Universidad de Sevilla

    Sevilla, España

    ROR https://ror.org/03yxnpp24

Book:
CAEPIA'97: actas
  1. Botti, Vicent (coord.)

Publisher: Vicent Botti ; Asociación Española para la Inteligencia Artificial (AEPIA)

ISBN: 978-84-8498-765-9,  84-8498-765-5

Year of publication: 1997

Pages: 489-498

Congress: Conferencia de la Asociación Española para la Inteligencia Artificial. (7. 1997. null)

Type: Conference paper

DIALNET GOOGLE SCHOLAR

Abstract

COGITO is a tool to obtain a classifier system from a labeled database. The result is a hierarchic rules system to organize the data in n-orthohedrons. This hierarchy means that the i-rule should be applied only it it does not satisfy the conditions of the previous (i-l)-rules, that is to say, in specific order to obtention. COGITO uses an evolutionary algorithm with real codification. The system can obtain various solutions by varying the parameters: number of rules and error rate. The computation time could be greater, but users would have more flexibility to choose the solution. At the end, we present the results of applying the system to several known databases, wich are quite good.

Data source: Dialnet