Modeling and Applying of RBF Neural Network Based on Fuzzy Clustering and Pseudo-Inverse Method

The key of advancing radial basis function neural network (RBFNN) is how to choose the data center and the number of cluster perfectly. In this paper fuzzy C-means clustering is used as K-means clustering ameliorated algorithm to determine the data center of RBF neural networks hidden nodes. Combined pseudo-inverse method RBF network model is constructed. Simulation test on the dimension predicting in selective laser sintering process showed the model is able to provide higher accurate predict, as well as less calculate quantity and quick training.