Exploring constructive algorithms with stopping criteria to produce accurate and diverse individual neural networks in an ensemble

Proceedings of the IEEE International Conference on Systems, Man and Cybernetics

Volumn 3, Issue , 2001, Pages 1526-1531

  Monirul Islam, M ^a   Shahjahan, M ^a   Murase, K ^a  

^a UNIVERSITY OF FUKUI   (Japan)

Author keywords

[No Author keywords available]

Indexed keywords

BENCHMARKING; CORRELATION METHODS; LEARNING ALGORITHMS; TUMORS;

CONSTRUCTIVE NEURAL NETWORK ENSEMBLE (CNNE);

NEURAL NETWORKS;

EID: 0035723264     PISSN: 08843627     EISSN: None     Source Type: Journal    
DOI: 10.1109/ICSMC.2001.973500     Document Type: Article

References (15)

1. Dynamic node creation in backpropagation networks
2. Temporal evolution of generalization during learning in linear networks
3. Bagging predictors
4. A decision theoretic generalization of on-line learning and an application to boosting
5. Neural network ensembles
6. Hierarchical mixtures of experts and the EM algorithm
7. Modeling with constructive backpropagation
8. Ensemble learning via negative correlation
9. Simultaneous training of negatively correlated neural networks in an ensemble
10. Advanced supervised learning in multilayer perceptrons-from back propagation to adaptive learning algorithm
11. Ensemble learning using decorrelated neural networks
12. The strength of weak learnability
13. Use of Quasi-Newton method in a feedforward neural network construction algorithm
14. Error correlation and error reduction in ensemble classifiers
15. A new evolutionary system for evolving artificial neural networks