A study on prediction model based on support vector regression for green technology automotive form design

Due to the air pollution and energy crisis, the added values to environmental protection from the green technology passenger cars have received scrutiny by consumers. In order to enhance the comprehension of consumers’ acceptance in green technology passenger cars, the goal of this study is to promote automotive designer’s understanding on the affective response of consumers on automotive form design. In general, consumers’ preference is mainly based on the vehicles’ form features that are traditionally manipulated by designers’ intuitive experience rather than an effective and systematic analysis. Therefore, when encountered the increasing competition in automotive market nowadays, enhancing the car designer’s understanding of consumers’ preference on the form features of green technology passenger cars to fulfill customers’ demands has become a common objective among automotive makers. In this paper, adjective evaluation data of customers were screened first from questionnaire to obtained baseline information. Secondly, automotive style features were systematically examined by numerical definition-based form representation. Finally, a predictive model based on these “adjectives” selected earlier was constructed using support vector regression (SVR) to incorporate the relationship between customers’ affective responses and automotive style features. The experimental results can be used for the future development of automotive, especially work as references of green technology passenger cars form design and support automotive makers to bring visual expectation on automotive form design to consumers’ experience. INTRODUCTION A product appearance plays an essential factor affecting consumers’ intention for purchasing. Therefore, the crux of developing a successful marketing strategy for promoting car sales is considered not only to fulfill customers’ demand on automobile functional utilities but to meet customers’ psychological needs on vehicles’ visual appearance [7]. In general, consumers’ preference is mainly based on the vehicles’ form features that are traditionally manipulated by designers’ intuitive experience rather than an effective and systematic analysis. The problem is that there is a shortage of reliable indicators to guide automobile form design to meet customers’ assess in automobile market. In order to access the guide of customers’ opinion and promote designers’ further understanding on customers’ psychological needs to a product, many studies regarding effective emotional design were proposed to get a better insight into customers’ subjective values [1, 5, 9]. Kansei engineering is a significant research based on a systematic investigation to inference customers’ perception to a satisfied product [3, 4, 8]. The technique also has been successfully used in car design [6, 12]. The main focuses of Kansei engineering is to examine and clarify the existing cause and effect relationship between affective factors for advanced enhancing product impression values [3]. Therefore, a Kansei model developed for predicting product form design is also regarded as an evaluation of regression problem. Various methodologies based on Kansei engineering technique have been presented to construct prediction model such as multiple linear regression (MLR), quantification theory type I (QT1), neural networks (NN), fuzzy logic, genetic algorithms (GA), and rough sets [11]. Moreover, Yang & Shieh proposed a support vector regression (SVR) to build a model for predicting customers’ feelings on mobile phones with very satisfactory performance.