Kansei Engineering-based Research On Design Optimization Of Banner Advertisement

With the promotion of online advertising in network applications, the spread effect of online advertising arouses more and more attention. Whether it can attract the user’s attention and achieve the expected spread effect or not, it is a very important issue that each advertiser is concerned about. The evaluation about Ad. effect is a major approach to help advertisers determine their advertising investment is worth or not. In the evaluation about advertising effectiveness, spread effectiveness is the core of it and the basis of economic effectiveness. And the advertising design is a key factor to impact the spread of advertising and the expected results.This study applies "Kansei engineering" for banner Ads.,which is the design support technology to connect product shape with image perception. The reason to use this theory is that most of the current advertising design is based on the designer’s understanding of advertising effects and few advertising is analyzed and designed from the perspective of browsers. Thus, it lacks the method of sense evaluation and how to design the banner Ads. according to the user’s perceptive needs.In this paper, Kansei Engineering is applied for banner Ads. design from the theory and experiment. It aims to find out the key design variables and their parameters for different image perceptions.The categories and items of banner Ads. design variables is determined by classified typical samples based on web search and expert opinions. Thirty-eight banner Ad. Prototype is designed with Flash CS5.First,10representative models are selected with the theory of multi-dimensional scale analysis and cluster analysis. Then scale of assessment for banner Ad. is built based on the representative image words. Next, the accuracy of appraisal scale is tested through coefficient analysis, factor analysis, independent samples T-Test and reliability&validity statistics. After that, the model between image perceptions and user’s preference is established with the method of BP neural network., while the model between design variables and image perceptions, user’s preference is established using the method of Quantification Type I. Finally, optimal solution is obtained through genetic algorithms approximately. At the same time, the optimal prototype is created with the help of Flash CS5.0and completed the validation.This study provides the design guidelines and method of banner Ad. with Kansei Engineering. It is a useful supplement for Human Factors Engineering design theory and also plays an important scientific significance. In addition, this study will bring the reference value for banner Ad. design and spread effectiveness. It has practical value and broad application prospects.