The Algorithm Of Clustering Based RBF-LBF NN And Its Application

The Radial Basis Function (RBF) method is a technology of interpolation in a high-dimensicnal space,which is at developing stage used to the neural network (NN).A RBFNN is a kind of feed forward network which basically involves 3 layers.It is proved that its speed of convergence of is much quicker than general BP algorithm. And its structure can be fixed in to algorithm. Therefore the study and application of RBFNN is getting fast development in recent years.At present,RBF NN had many learning algorithms, commonly used k-means and orthogonal least squares method training hidden layer parameters,used gradient descent algorithm and recursive least squares method training weight.The main problems of the design of these algorithms include the number of nodes in hidden layer, the center and the radius determination,and the training of network weights.This dissertation has mainly launched the following work to the above-mentioned problems.Firstly,this paper discussed the k-means clustering algorithm of the advantages and disadvantages, and proposed a Self-adaptive k-means clustering algorithm (SA-K-means).And then this paper studied single-layer neural network and RBF network classification of the principle of LBF and RBF-LBF neural network in series classification and proposed SA-K-means clustering based RBF-LBF neural networks tandem learning algorithm (CBRBF-LBF). The algorithm used clustering in the single-layer RBF network training in order to determine the initial structure of the network, and then by adjusting the sub-sample of the type of mistake, so that overlap or merging of nuclear function.Experiments show that the CBRBF-LBF algorithm can rapidly reaching a stable network structure and network and achieve a smaller kernel value. At the same time, this paper combine the case of mathematical expression recognition, proposed a baseline based multi-candidate mathematical expression recongnition method, and used CBRBF-LBF neural network in the case of mathematical expression character recognition,achieved good results.