Dynamic mapping of design elements and affective responses: a machine learning based method for affective design

ABSTRACT

Affective design has received more and more attention. Kansei engineering is widely used to transform consumers’ affective needs into product design. Yet many previous studies used questionnaire survey to obtain consumers’ affective responses, which is usually in a small scale, not updated, time-consuming and labour-intensive. The life cycle of a product is getting shorter and shorter, social trends are changing unconsciously, which results in the change of consumers’ affective responses as well. Therefore, it’s necessary to develop an approach for collecting consumers’ affective responses extensively, dynamically and automatically. In this paper, a machine learning-based affective design dynamic mapping approach (MLADM) is proposed to overcome those challenges. It collects consumers’ affective responses extensively. Besides, the collection process is continuous because new users can express their affective responses through online questionnaire. The products information is captured from online shopping websites and the products’ features and images are extracted to generate questionnaire automatically. The data obtained are utilised to establish the relationship between design elements and consumers’ affective responses. Four machine learning algorithms are used to model the relationship between design elements and consumers’ affective responses. A case study of smart watch is conducted to illustrate the proposed approach and validate its effectiveness.

KEYWORDS:

• Affective design
• Kansei engineering
• machine learning
• affective responses
• design elements

Acknowledgement

This work was supported by the National Natural Science Foundation of China under Grant [51405089]; and the Science and Technology Planning Project of Guangdong Province under [2015B010131008] and [2015B090921007].

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

W. M. Wang http://orcid.org/0000-0003-1188-4537

G. Q. Huang http://orcid.org/0000-0003-3592-7917

Additional information

Funding

This work was supported by the National Natural Science Foundation of China under [grant no 51405089]; and the Science and Technology Planning Project of Guangdong Province under [grant no 2015B010131008 and 2015B090921007].