A novel data-driven method for product aesthetics evaluating and optimising based on knowledge graph

With increasing competition from economic globalisation, many manufacturing companies strive to maintain their competitive advantage by meeting users' personalised requirements. Hence, numerous enterprises provided personalised products by inviting users to complete product configurations online. Most users have not received systematic training, so their configured solutions may not comply with the essential aesthetics principle. Meanwhile, the design teams must ensure that the configured products conform to essential aesthetics to maintain the brand reputation. However, most existing studies were proposed for the single-product design solution, which cannot assist designers in evaluating and optimising mass user-configured solutions efficiently and automatically. It still lacks a new method to evaluate and optimise user-configured product solutions automatically from the essential aesthetic perspective. To fill this gap, a novel data-driven method is proposed to select the proper product observing perspective, conduct the comprehensive aesthetic evaluation based on the general traditional aesthetic principles, and propose optimisation suggestions based on the knowledge graph. To validate the feasibility of the proposed method, the bicycle is taken as the case in this paper.