The identification of the proper structure of nonlinear neural networks(NNs) is a difficult problem, since these black-box models are not interpretable. Theaim of the paper is to propose a new approach that can be used for the analysis andthe reduction of these models. It is shown that NNs with sigmoid transfer function canbe transformed into fuzzy systems. Hence, with the use of this transformation NNscan be analyzed by human experts based on the extracted linguistic rules. Moreover,based on the similarity of the resulted membership functions the hidden neurons ofthe NNs can be mapped into a two dimensional space. The resulted map providesan easily interpretable figure about the redundancy of the neurons. Furthermore, thecontribution of these neurons can be measured by orthogonal least squares techniquethat can be used for the ordering of the extracted fuzzy rules based on their importance.A practical example related to the dynamic modeling of a chemical process system isused to prove that synergistic combination of model transformation, visualization andreduction of NNs is an effective technique, that can be used for the structural andparametrical analysis of NNs.
展开▼
