A New Heuristic for Radius Estimation in RBF Neural Networks
Resumo
Redes Neurais baseadas em Funções de Base Radial (RBFNN) são métodos clássicos do aprendizado de máquina que contêm uma camada de Funções de Base Radial (RBF) que atuam como extrator de características para a camada final, que executa o reconhecimento de padrões. A estimação do raio das RBFs é uma das atividades mais cruciais do treinamento de modelos RBFNN e afeta diretamente o seu poder de generalização e acurácia. Neste trabalho é apresentado uma nova heurística para estimação do raio e experimentos computacionais são empregados para medir sua eficácia comparada à outras abordagens usando 14 problemas de classificação. A método proposta mostrou uma eficácia competitiva, vencendo os demais métodos em 9 dos 14 problemas.
Referências
Alves, M. A. D., Pinto, J. O. P., Galotto, L., Kimpara, M. L. M., Garcia, R. C., Godoy, R. B., Teixeira, H. C. G., and Campos, M. C. M. (2021). A modified algorithm for training and optimize rbf neural networks applied to sensor measurements validation. IEEE Sensors Journal.
Barreto, G. A., Brasil, I. S., and Souza, L. G. M. (2021). Revisiting the modeling of wind turbine power curves using neural networks and fuzzy models: an application-oriented evaluation. Energy Systems, pages 1–28.
BATISTA, P. V., BRAGA, A. D. P., and TORRES, L. C. B. (2019). Cálculo do raio em In Congresso Brasileiro de redes rbf baseado na distância média entre os centros. Automática-CBA.
Benoudjit, N., Archambeau, C., Lendasse, A., Lee, J. A., Verleysen, M., et al. (2002). Width optimization of the gaussian kernels in radial basis function networks. In ESANN, volume 2, pages 425–432.
Bodyanskiy, Y., Pirus, A., and Deineko, A. (2020). Multilayer radial-basis function netIn 2020 IEEE 15th International Conference on Computer work and its learning. Sciences and Information Technologies (CSIT), volume 1, pages 92–95. IEEE.
Braga, A. d. P., Carvalho, A., and Ludermir, T. B. (2000). Redes neurais artificiais: teoria e aplicações. Livros Técnicos e Científicos.
Buhmann, M. D. (2003). Radial basis functions: theory and implementations, volume 12. Cambridge university press.
Burruss, M., Ramakrishna, S., and Dubey, A. (2021). Deep-rbf networks for anomaly detection in automotive cyber-physical systems. arXiv preprint arXiv:2103.14172.
Dong, C.-R., Chan, P. P. K., Ng, W. W. Y., and Yeung, D. S. (2011). A survey of the initialization of centers and widths in radial basis function network for classification. In 2011 International Conference on Machine Learning and Cybernetics, volume 3, pages 1082–1087.
Dua, D. and Graff, C. (2017). UCI machine learning repository.
Han, Z., Qian, X., Huang, H., and Huang, T. (2021). Efficient design of multicolumn rbf networks. Neurocomputing, 450:253–263.
Haykin, S. (1994). Neural networks: A comprehensive foundation: Macmillan college publishing company. New York.
Hou, Y. (2018). KEEL dataset repository. ZSCC: NoCitationData[s0] Type: dataset.
Kaaniche, K., Rashid, N., Miraoui, I., Mekki, H., and El-Hamrawy, O. I. (2021). Mobile robot control based on 2d visual servoing: A new approach combining neural network with variable structure and atness theory. IEEE Access, 9:83688–83694.
Rumelhart, D. E. (1986). Learning internal represectations by error propagation. Parallel distributed processing, 1(8).
Verleysen, M. and Hlavackova, K. (1994). An optimized rbf network for approximation of functions. In ESANN’1994 proceedings European Symposium on Artificial Neural Networks.
Wang, C., Cui, L., Liang, M., Li, J., and Wang, Y. (2021). Adaptive neural network control for a class of fractional-order nonstrict-feedback nonlinear systems with fullIEEE Transactions on Neural Networks and state constraints and input saturation. Learning Systems, pages 1–13.
Wu, H., Zhao, Y.-P., Yang, T.-L., and Tan, H.-J. (2021). An ensemble radius basis function network based on dynamic time warping for real-time monitoring of supersonic inlet ow patterns. Aerospace Science and Technology, 111:106551.
Zadeh, P. H., Hosseini, R., and Sra, S. (2018). Deep-rbf networks revisited: Robust classification with rejection. arXiv preprint arXiv:1812.03190.
Zhao-zhao, Z., Qiu-wan, W., and Ying-qin, Z. (2020). A self-adaptive multi-hierarchical modular neural network for complex problems. In International Conference on Verification and Evaluation of Computer and Communication Systems, pages 244–256. Springer.
Barreto, G. A., Brasil, I. S., and Souza, L. G. M. (2021). Revisiting the modeling of wind turbine power curves using neural networks and fuzzy models: an application-oriented evaluation. Energy Systems, pages 1–28.
BATISTA, P. V., BRAGA, A. D. P., and TORRES, L. C. B. (2019). Cálculo do raio em In Congresso Brasileiro de redes rbf baseado na distância média entre os centros. Automática-CBA.
Benoudjit, N., Archambeau, C., Lendasse, A., Lee, J. A., Verleysen, M., et al. (2002). Width optimization of the gaussian kernels in radial basis function networks. In ESANN, volume 2, pages 425–432.
Bodyanskiy, Y., Pirus, A., and Deineko, A. (2020). Multilayer radial-basis function netIn 2020 IEEE 15th International Conference on Computer work and its learning. Sciences and Information Technologies (CSIT), volume 1, pages 92–95. IEEE.
Braga, A. d. P., Carvalho, A., and Ludermir, T. B. (2000). Redes neurais artificiais: teoria e aplicações. Livros Técnicos e Científicos.
Buhmann, M. D. (2003). Radial basis functions: theory and implementations, volume 12. Cambridge university press.
Burruss, M., Ramakrishna, S., and Dubey, A. (2021). Deep-rbf networks for anomaly detection in automotive cyber-physical systems. arXiv preprint arXiv:2103.14172.
Dong, C.-R., Chan, P. P. K., Ng, W. W. Y., and Yeung, D. S. (2011). A survey of the initialization of centers and widths in radial basis function network for classification. In 2011 International Conference on Machine Learning and Cybernetics, volume 3, pages 1082–1087.
Dua, D. and Graff, C. (2017). UCI machine learning repository.
Han, Z., Qian, X., Huang, H., and Huang, T. (2021). Efficient design of multicolumn rbf networks. Neurocomputing, 450:253–263.
Haykin, S. (1994). Neural networks: A comprehensive foundation: Macmillan college publishing company. New York.
Hou, Y. (2018). KEEL dataset repository. ZSCC: NoCitationData[s0] Type: dataset.
Kaaniche, K., Rashid, N., Miraoui, I., Mekki, H., and El-Hamrawy, O. I. (2021). Mobile robot control based on 2d visual servoing: A new approach combining neural network with variable structure and atness theory. IEEE Access, 9:83688–83694.
Rumelhart, D. E. (1986). Learning internal represectations by error propagation. Parallel distributed processing, 1(8).
Verleysen, M. and Hlavackova, K. (1994). An optimized rbf network for approximation of functions. In ESANN’1994 proceedings European Symposium on Artificial Neural Networks.
Wang, C., Cui, L., Liang, M., Li, J., and Wang, Y. (2021). Adaptive neural network control for a class of fractional-order nonstrict-feedback nonlinear systems with fullIEEE Transactions on Neural Networks and state constraints and input saturation. Learning Systems, pages 1–13.
Wu, H., Zhao, Y.-P., Yang, T.-L., and Tan, H.-J. (2021). An ensemble radius basis function network based on dynamic time warping for real-time monitoring of supersonic inlet ow patterns. Aerospace Science and Technology, 111:106551.
Zadeh, P. H., Hosseini, R., and Sra, S. (2018). Deep-rbf networks revisited: Robust classification with rejection. arXiv preprint arXiv:1812.03190.
Zhao-zhao, Z., Qiu-wan, W., and Ying-qin, Z. (2020). A self-adaptive multi-hierarchical modular neural network for complex problems. In International Conference on Verification and Evaluation of Computer and Communication Systems, pages 244–256. Springer.
Publicado
29/11/2021
Como Citar
SOUZA, Lucas Pimenta de; BATISTA, Paulo V. C.; SILVA, Petrônio C. L..
A New Heuristic for Radius Estimation in RBF Neural Networks. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 18. , 2021, Evento Online.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2021
.
p. 189-196.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2021.18252.
