FINDING PROTOTYPES FOR NEAREST-NEIGHBOR CLASSIFICATION BY MEANS OF GRADIENT DESCENT AND DETERMINISTIC ANNEALING

Authors
DECAESTECKER C
Citation
C. Decaestecker, FINDING PROTOTYPES FOR NEAREST-NEIGHBOR CLASSIFICATION BY MEANS OF GRADIENT DESCENT AND DETERMINISTIC ANNEALING, Pattern recognition, 30(2), 1997, pp. 281-288
Citations number
20
Categorie Soggetti
Computer Sciences, Special Topics","Engineering, Eletrical & Electronic","Computer Science Artificial Intelligence
Journal title
Pattern recognition → ACNP
ISSN journal
00313203
Volume
30
Issue
2
Year of publication
1997
Pages
281 - 288
Database
ISI
SICI code
0031-3203(1997)30:2<281:FPFNCB>2.0.ZU;2-Q
Abstract
A new method is presented to find prototypes for a nearest neighbour c lassifier. The prototype locations are optimised through a gradient de scent and a deterministic annealing process. The proposed algorithm al so includes an initialisation strategy which aims to provide the maxim um classification rate on the training set with the minimum number of prototypes. Experiments show the efficiency of this algorithm on both real and artificial data. Copyright (C) 1997 Pattern Recognition Socie ty.