Title :
A Unified Strategy to Deal with Different Natures of Reject
Author :
Mouchère, Harold ; Anquetil, Eric
Author_Institution :
IRISA, Rennes
Abstract :
The interest of reject for classifier optimization has been shown many times. The diversity of the applications requiring this concept makes us to distinguish two main natures of reject with distinct goals: the confusion reject and the distance reject. After the description of this two kinds of reject, we present a unified formalism to define them using reliability functions and reject thresholds. Then we present a generic algorithm dedicated to the automatic learning of these thresholds. Finally, we compare various possibilities of reject to achieve application goals
Keywords :
genetic algorithms; learning (artificial intelligence); pattern classification; automatic learning; classifier optimization; confusion reject; distance reject; generic algorithm; reject thresholds; reliability functions; unified formalism; Constraint optimization; Fuzzy neural networks; Fuzzy systems; Histograms; Particle swarm optimization; Pattern recognition; Prototypes; Radial basis function networks;
Conference_Titel :
Pattern Recognition, 2006. ICPR 2006. 18th International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
0-7695-2521-0
DOI :
10.1109/ICPR.2006.193
