Dynamics of Continuous, Discrete and Impulsive Systems: Series B, Applications and Algorithms
Volume 14, Supplementary Issue 2, June 2007, Pages 1598-1601

A neural network approach to website design

^aClayton School of Information Technology, Faculty of Information Technology, Monash University, Clayton, Victoria 3800, Australia

^bDepartment of Arts and Design, College of Arts, National Hualien University of Education, Hualien, 970, Taiwan

^cDepartment of Transportation and Communications Management Science, National Cheng Kung University, Tainan, 701 Taiwan

Abstract

This paper develops a neural network (NN) approach for examining how specific design elements of a website will affect users' perception of the website characterized by image words. An experimental study on 96 university websites is conducted based on the process of Kansei engineering. With a morphological analysis and a cluster analysis, the study identifies 7 influential website design elements and 33 representative websites as experimental samples for training and testing NN models. Four NN models are built to formulate the relationship between the 7 website design elements and three website image words including easy-difficult, clear-confusing, and realistic-fancy. The results of the NN models can be used to construct a website design database for determining the best combination of website design elements for matching a set of desirable website images as perceived by users. The result of this study provides useful insights in supporting the design of website elements that better meet users' specific expectations.

Indexed Terms: Neural network applications, user centered design, Kansei engineering, website design.

Corresponding author. Tel.: +61 3 9905 5808; fax: +61 3 9905 5159. Email: ChungHsing.Yeh@infotech.monash.edu.au