Research on Visual Performance Evaluation Model of E-commerce Websites

With the growth of users' personalized demand for c-commerce visual design and the rapid development of machine learning technology, a large number of Internet pages are automatically designed and generated by computers. :After being generated, effective evaluation mechanisms are needed for verification and feedback At present, the evaluation model of visual design generally takes traditional web pages as the research object, and there is little research on the complexity and marketing characteristics of e-commerce information. It is necessary to consider how visual design can promote the effectiveness of communication between users and web pages from the perspectives of psychology , behavior and content. This paper proposes a "multi-dimensional visual performance evaluation model" for the evaluation after the generation of contemporary ecommerce web page design. This paper first constructs the visual marketing system and discusses the circulation mechanism of user demand import, e-commerce web page design, multi-dimensional visual performance evaluation and visual precision marketing. Under the system, a relational model of how "visual performance indicators (aesthetic performance, search behavior performance, information usefulness performance)" based on "scene factors" affect "overall evaluation" is proposed, defined as a multidimensional visual performance evaluation model, and verified by quantitative experiments. The model provides visual design basis for designers and recyclable evaluation data for machine learning generated pages.