|Title:||Optimization of radial basis function networks for pattern classification|
Neural networks (Computer science)
Hong Kong Polytechnic University -- Dissertations
|Department:||Department of Electronic Engineering|
|Pages:||63 leaves : ill. ; 30 cm|
|Abstract:||Radial basis function (RBF) networks were introduced by Broomhead and Lowe (1988)  and are known to be universal approximators. These networks approximate an unknown function from sample data by positioning radially symmetric "localized receptive fields" over portions of the input space that contain the sample data. Due to the local nature of the network, standard clustering algorithm such as K-means clustering are often used to determine the centers of the receptive fields. In this thesis, we analyse how an RBF network trained with the backpropagation (BP) learning algorithm can be applied to pattern classification problems. We also present a new algorithm for the construction of RBF networks with optimal number of hidden units. Thereby providing a solution to the problem of determining the number of hidden units and the widths required to get a good network solution. In this approach, we use the total mean squared error (MSE) information to determine when to generate and insert a new hidden unit to an RBF network. Then it was tested by the exclusive-or and the two intertwined spiral problems in order to evaluate the classification capability of the network. The results show that the algorithm was able to generate an appropriate number of hidden units and to provide a smooth fit to the training data.|
|Rights:||All rights reserved|
Files in This Item:
|b12197981.pdf||For All Users (off-campus access for PolyU Staff & Students only)||2.18 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: