Citation

Goh, Pey Yun (2018) Reduced Radial Basis Function Neural Networks For Data Classification. PhD thesis, Multimedia University.

Full text not available from this repository.

Abstract

The research in this thesis is related to design and develop three models of radial basis function network (RBFN) that handle spurious neurons among hidden neurons. A two-stage learning algorithm of RBFN is presented. Dynamic decay adjustment (DDA) is used in the first stage. This forms the radial basis function network which trained with the DDA (i.e. RBFNDDA). DDA enables the RBFN to learn incrementally without pre-defining the number of hidden neurons. It is fast with this greedy insertion behaviour but spurious neurons are included during the learning process. This behaviour leads to unnecessary huge network architecture due to spurious neurons which include redundant and outlier neurons. Prediction capability of RBFNDDA may be affected due to these spurious neurons too. Due to this problem, the second stage is exclusively devoted to the greedy insertion investigation of RBFNDDA with concern to two issues, i.e. redundant and outlier neurons. Two different techniques are proposed to explore the redundant and outlier neurons. A modified histogram algorithm is proposed to prune the redundant neurons.

Item Type:	Thesis (PhD)
Additional Information:	Call No.: QA76.87 .G64 2018
Uncontrolled Keywords:	Neural networks (Computer science)
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science > QA76.75-76.765 Computer software
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	30 Aug 2021 10:41
Last Modified:	25 Jun 2026 08:13
URII:	http://shdl.mmu.edu.my/id/eprint/9450

Downloads

Edit (login required)