Data Preprocessing in Multilayer Perceptron Modelling
Ku R. Ku-Mahamud*, Aniza M. Din and Fathilah M. Alipiah* Author for corresponding; e-mail address: ruhana@uum.edu.my
Volume: Vol.29 No.1 (APRIL 2002)
Opinion
DOI:
Received: 13 Febuary 2002, Revised: -, Accepted: 2 April 2002, Published: -
Citation: Ku-mahamud K.R., Din A.M. and Alipiah F.M., Data Preprocessing in Multilayer Perceptron Modelling, Chiang Mai Journal of Science, 2002; 29(1): 43-48.
Abstract
Obtaining a good multilayer perceptron prediction model depends on the choice of data preprocessing techniques. In this study, data were preprocessed, represented and scaled, using two data representation techniques combined with four data scaling techniques. The prediction capability of the network can be enhanced using the proposed techniques.