Density based grid clustering partition of the input space for RBF neural network

Authors:
Ping Wei;Wenbo Liu;Xin Zuo;Xionglin Luo
Affiliations:
Department of Automation, China University of Petroleum, Beijing, P.R. China;Department of Automation, China University of Petroleum, Beijing, P.R. China;Department of Automation, China University of Petroleum, Beijing, P.R. China;Department of Automation, China University of Petroleum, Beijing, P.R. China
Venue:
AICI'12 Proceedings of the 4th international conference on Artificial Intelligence and Computational Intelligence
Year:
2012

Citing 6
Cited 0

Radial basis function network configuration using genetic algorithms

Neural Networks
Automatic subspace clustering of high dimensional data for data mining applications

SIGMOD '98 Proceedings of the 1998 ACM SIGMOD international conference on Management of data
Fast learning in networks of locally-tuned processing units

Neural Computation
A Radial Basis Function network training algorithm using a non-symmetric partition of the input space - Application to a Model Predictive Control configuration

Advances in Engineering Software
Comparing support vector machines with Gaussian kernels to radialbasis function classifiers

IEEE Transactions on Signal Processing
Orthogonal least squares learning algorithm for radial basis function networks

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Selecting proper centers is important for constructing a radial basis function (RBF) neural network. Motivated by the idea of clustering according to density for data mining applications, the algorithm of unevenly partition of the input space is proposed in the paper. By combining the neighboring subsets with low density data, the ultimate clustering centers are selected as the hidden layer centers in RBF neural network. An example is presented to demonstrate the method proposed, and the results illustrate the comparative high accuracy RBF neural network with comparative short training time can be created by selecting the initial partition set and the upper limit for the number of data in one subinterval properly.