Towards a global method for comparing classifiers
- Cabaña, Alejandra
- Gamboa, Fabrice
Editorial: Comité organizador del XXX Congreso Nacional de Estadística e Investigación Operativa y IV Jornadas de Estadística Pública
ISBN: 978-84-690-7249-3
Año de publicación: 2007
Congreso: Congreso Nacional de Estadística e Investigación Operativa (30. 2007. Valladolid)
Tipo: Aportación congreso
Resumen
It is well known, from empirical comparisons, that dierent algorithms show dierent performances when applied to dierent data sets. Our aim is to provide an universal instrument for choosing among I classiers, when nothing is known a priori about the structure of data set to be classied. In this case, it might not suce to look just at classication mean errors: it would be advisable to use a classication algorithm with low variance.