Conference Proceeding

Training Classifiers for Tree-Structured Sets of Categories

Dept. of Signal Theor. & Commun., Univ. Carlos III de Madrid
10/2005; DOI:10.1109/MLSP.2005.1532916 ISBN: 0-7803-9517-4 pp.291 - 296 In proceeding of: Machine Learning for Signal Processing, 2005 IEEE Workshop on
Source: IEEE Xplore

ABSTRACT In this paper we propose a new method for training classifiers for multi-class problems when classes are not (necessarily) mutually exclusive and may be related by means of a probabilistic tree structure. Our method is based on the definition of a Bayesian model relating network parameters, feature vectors and categories. Learning is stated as a maximum likelihood estimation problem of the classifier parameters. The proposed algorithm is tested on an image retrieval scenario

0 0
 · 
0 Bookmarks
 · 
22 Views
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.

Keywords

Bayesian model
 
classes
 
image retrieval scenario
 
new method
 
probabilistic tree structure
 
proposed algorithm
 
training classifiers