Improved nonlinear PCA based on RBF networks and principal curves

  • Authors:

  • Xueqin Liu;Kang Li;Marion McAfee;Jing Deng

  • Affiliations:

  • School of Electronics, Electrical Engineering and Computer Science, Queen's University, Belfast, UK;School of Electronics, Electrical Engineering and Computer Science, Queen's University, Belfast, UK;Department of Mechanical and Electronic Engineering, Institute of Technology Sligo, Sligo, Ireland;School of Electronics, Electrical Engineering and Computer Science, Queen's University, Belfast, UK

  • Venue:
  • LSMS/ICSEE'10 Proceedings of the 2010 international conference on Life system modeling and and intelligent computing, and 2010 international conference on Intelligent computing for sustainable energy and environment: Part I
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Nonlinear PCA based on neural networks (NN) have been widely used in different applications in the past decade. There is a difficulty with the determination of the optimal topology for the networks that are used. Principal curves were introduced to nonlinear PCA to separate the original complex five-layer NN into two three-layer RBF networks and eased the above problem. Using the advantage of Fast Recursive Algorithm, where the number of neurons, the location of centers, and the weights between the hidden layer and the output layer can be identified simultaneously for the RBF networks, the topology problem for the nonlinear PCA based on NN can thus be solved. The simulation result shows that the method is excellent for solving nonlinear principal component problems.