Application of a radial basis function neural network for diagnosis of diabetes mellitus

摘要: In this article an attempt is made to study the appl i- cability of a general purpose, supervised feed forward neural network with one hidden layer, namely. Radial Basis Function (RBF) network. It uses rela- tively smaller number locally tuned units and adaptive in nature. RBFs are suitable for pattern re cog- nition classification. Performance RBF was also compared most commonly used multilayer perceptron model classical logistic regression. Diabetes database empirical comparisons results show that performs better than other models. 1 . MLP, weighted sum inputs bias term passed activation level through transfer function produce output, arranged layered feed-forward topology called Feed Forward Neural Network (FFNN). The schematic representation FFNN n inputs, m output unit along input given Figure 1. An artificial ne twork (ANN) has three layers: layer layer. vastly increases lear ning power MLP. or modifies give desired output. chosen such algorithm requires r e- sponse continuous, single-valued first derivative existence. Choice hi d- den layers, nodes type activ ation plays important role constructions 2-4 basis based on learning. networks were independently proposed by many researchers 5-9 popular alter- native good at modelling nonlinear data can be trained stage rather using iterative process as MLP learn application quickly. They useful sol ving problems where corrupted add itive noise. transformation functions Gau ssian distribution. If error minimized appropr ately, it will outputs unity, which represent probability outputs. objective applic ability diabetes compare gression.