A Basic Study of Method Dealing with Non-Linearity of Kansei

感性工学是大约30年前由广岛大学的长池光雄教授创立的，它已经在日本和世界各地传播开来。它被定义为“将消费者对即将到来的产品的感觉和形象(感性)转化到设计领域的技术”。由于在产品设计和功能上贯彻了以消费者为中心的感性思维，通过感性思维工程开发的新产品非常畅销，有时还在市场上构筑了新的弯道。到目前为止，感性工学已经在汽车、工程机械、家电、房屋建筑、服装、化妆品等行业被引入，据说感性工学总体上具有非线性特征。然而，在统计分析方面，我们将其作为线性处理来处理。本研究所关注的是弄清感性思维是否具有线性或非线性，并构建能够加热感性思维线性特征的方法。为此而进行的实验表明，感性文字的线性依赖于感性文字本身。例如，与身体特征有关的感性词语，如“大-小”、“重-轻”、“简单-复杂”等，在评价数据中表现出线性。但那些与审美有关的词语，如“美-丑”、“美-不美”等，则导致了这种非线性特征。我们通过因子分析、量化理论和非线性回归分析得出了这些有趣的结果。

Kansei Engineering was founded by Professor Mitsuo Nagamachi at Hiroshima University about 30 years ago and it has spread out in Japan as well as in the world. It is defined as "a translating technology of consumers' feeling and image (Kansei) on a coming product into the design domain". Since the consumers' kansei is implemented in the product design and its function, the new products developed through the kansei engineering have sold very well and sometime it constructed the new bend in the market. The kansei engineering has been introduced so far in industries of Automobile, Construction Machine, Electric Home Appliance, House Construction, Costume, Cosmetics and so forth.It is said that the kansei has in general the non-linear characteristics. However, we have dealt with it as linearity for treatment in terms of statistical analysis. The concern of the present research is to make clear whether the kansei has linearity or non-linearity and to construct the method, which is able to heat the linear characteristics of kansei. The experiments conducted for this purpose showed that the kansei linearity relies on the kansei words themselves. For instance, the kansei words related to the physical trait like "large-small", "heavy-light", "simple-complicated" and so on showed the linearity in the evaluated data. But those words concerned with aesthetic like "beautiful-ugly", "graceful-not graceful" and the like resulted in the non-linear feature. We found out these interesting results in terms of Factor Analysis, Quantification Theory and non-linear regression analysis.