Towards Developing Computational Models to Predict Perceived Visual Aesthetics of Website Interface Design

This study comes within a framework of scientific research arguing that users' perception of visual aesthetics of computer interface design is related to visual screen design features and layout elements. This framework is concerned with determining what features triggers users' perception of aesthetics and tries to express such features numerically. The goal is to combine their collective effects using mathematical formulas and computational models that would objectively predict perceived visual aesthetics.;The general purpose of this study is to continue the research efforts towards this goal. In this study, rigorous experimental methods were utilized to verify and further improve currently available measures and models. This included extending the application of these measures and models to the case of website interface design. Viability of this extension was assessed using standard questionnaires deigned to measure perceived visual aesthetics of website design.;Results of this study confirmed findings of previous studies regarding the use of visual features to predict perceived visual aesthetics and further support the concept of expressing such feature using mathematical formulas and use them in turn as basis to develop computational models to predict aesthetics. Results also proved that such screen layout-based measures can work with website interfaces. Moreover, results showed that objective screen layout-based measures relates to subjective questionnaire-based measure. The relationship is particularly stronger with questionnaire elements related to screen layout. This supports further the suggestion that objective layout-based measures could be used to generally assess the overall visual aesthetics of websites and particularly aesthetic aspects related to classical and simplicity dimensions of website aesthetics.