A performance evaluation of pruning effects on hybrid neural network

Abstract

In this paper, we explore the pruning effects on a hybrid mode sequential learning algorithm
namely FuzzyARTMAP-prunable Radial Basis Function (FAM-PRBF) that utilizes Fuzzy
ARTMAP to learn a training dataset and Radial Basis Function Network (RBFN) to perform
regression and classification. The pruning algorithm is used to optimize the hidden layer of
the RBFN. The experimental results show that FAM-PRBF has successfully reduced the
complexity and computation time of the neural network.

Keywords: pruning; radial basis function network; fuzzy ARTMAP.