The article aims to explore a better way of user needs understanding and a rapid prototype design in product iteration and development. Several textual sentiment analysis methods such as HowNet dictionary, BERT and KMeans was integrated into kansei engineering to recognize the kansei words more precisely. The design process proposed by the study was used in electric vehicle charging pile design. As an affiliated product in the car industry, the charging pile design iterates quickly, therefore a faster way to understand customers with less resource consumption is needed. By using the core HowNet dictionary, the study established a kansei word dictionary from online user reviews and articles. The user reviews and articles were also processed by BERT pretrained model and clustered by KMeans to show more information of the text, in case of some key words missing when using HowNet independently. Then, 10 kansei word pairs were defined to represent the user’ sentiment towards the existing charging piles. Finally, semantic differential method was used to evaluate the existing piles design to scope the users’ preferences and lead the prototype design. Following the preferences of users, three prototypes were designed to validate if the sentiment of users were correctly understood. The study shown a rapid design process that is effective in the development of products like charging piles. The feasibility and accuracy in Chinese kansei word extraction by HowNet dictionary were also validated.
Research on Electric Vehicle Charging Pile Design Based on Kansei Engineering and Textual Sentiment Analysis
Lect.Notes Computer
International Conference on Human-Computer Interaction ; 2024 ; Washington DC, USA June 29, 2024 - July 04, 2024
01.06.2024
14 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Vehicle development through "Kansei" engineering
Kraftfahrwesen | 2003
|Vehicle Development through “Kansei” Engineering
SAE Technical Papers | 2003
|