Analysis of Specifications for Web Screen-Design Using Rough Sets

Abstract

Recent Web sites have created the impression of representing companies or universities. Little research has been done to verify the relations between the composition elements of Web sites and the sense of values using nonlinear methods such as rough sets.
We made a Web site for a university in the design and information design field into a case study, and aimed at clarifying the relation between the composition elements and users' sense of values regarding the Web site. First, a questionnaire for sample Web sites was conducted using six terms to evaluate the sense of values. We then analyzed the results with multiple regression analysis, and clarified what evaluation terms affected their degree of aesthetic appeal. As a result, the subjects were classified by their sense of values into five clusters. Furthermore, combinations of composition elements for the Web site affecting the degree of aesthetic appeal in each cluster were extracted using rough sets. As a result of assessing the new Web site on the basis of the above results, we found that impressions and aesthetic appeal are controllable.