On the Training of Artificial Neural Networks with Radial Basis Function Using Optimum-Path Forest Clustering

• Autor(a) principal: Rosa, Gustavo H. [UNESP]
• Data de Publicação: 2014
• Outros Autores: Costa, Kelton A. P. [UNESP], Passos Junior, Leandro A. [UNESP], Papa, Joao P. [UNESP], Falcao, Alexandre X., Tavares, Joao Manuel R. S., IEEE
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1109/ICPR.2014.262; http://hdl.handle.net/11449/186395
• topic: Artificial Neural Networks; Radial Basis Function; Optimum-Path Forest

In this paper, we show how to improve the Radial Basis Function Neural Networks effectiveness by using the Optimum-Path Forest clustering algorithm, since it computes the number of clusters on-the-fly, which can be very interesting for finding the Gaussians that cover the feature space. Some commonly used approaches for this task, such as the well-known k-means, require the number of classes/clusters previous its performance. Although the number of classes is known in supervised applications, the real number of clusters is extremely hard to figure out, since one class may be represented by more than one cluster. Experiments over 9 datasets together with statistical analysis have shown the suitability of OPF clustering for the RBF training step.