Quantitative evaluation of dependence among outputs in ECOC classifiers using mutual information based measures

Proceedings of the International Joint Conference on Neural Networks

Volumn 2, Issue , 2001, Pages 784-789

  Masulli, F ^a   Valentini, G ^a  

^a UNIVERSITY OF GENOA   (Italy)

Author keywords

[No Author keywords available]

Indexed keywords

CLASSIFICATION (OF INFORMATION); ENCODING (SYMBOLS); ERROR CORRECTION; IMAGE PROCESSING; LEARNING SYSTEMS; MATHEMATICAL TRANSFORMATIONS;

CODEWORD BITS; ERROR CORRECTING OUTPUT CODING CLASSIFICATION METHODS; INDEPENDENT NONLINEAR DICHOTOMIZERS; MUTUAL INFORMATION BASED MEASURES;

MULTILAYER NEURAL NETWORKS;

EID: 0034877853     PISSN: None     EISSN: None     Source Type: Conference Proceeding    
DOI: None     Document Type: Conference Paper

References (17)

1. An information-maximization approach to blind separation and blind deconvolution
2. Support vector networks
3. Solving multiclass learning problems via error-correcting output codes
4. Measures of diversity in classifier ensembles
5. Effectiveness of error correcting output codes in multiclass learning problems
6. Parallel non linear dichotomizers
7. Mutual information methods for evaluating dependence among outputs in learning machines
8. UCI repository of machine learning databases
9. Mutual information in learning feature transformations
10. A modular neural network for enhancement of cross-polar radar targets
11. NEURObjects, a set of library classes for neural networks development