Research On Optimal Design Of Web Visual Flow Based On User Behavior Mining

The purpose of this article is to use random forest decision trees and nearest neighbor methods to predict the probability distribution of the visual element competition results in different visual processes of the WEB interactive interface,and provide a basis for optimizing the layout of the WEB interactive interface.In the study,the main characteristics of the interface-jumping visual flow were discussed.The random-jumping visual flow characteristics of the WEB interactive interface were derived from its inherent visual competition mechanism.Conjectures were made on the influence factors of competition and correlation experiments were conducted to verify.The validated impact factors were trained and analyzed,and a decision tree and KNN model to simulate the competitive process were constructed.Through the interface optimization experiment results,using the quantitative indicators of user behavior to evaluate the optimization results,it is proved that using this idea can significantly improve the user's visual process,so that the visual flow can be carried out according to the design goals.In addition,the research experiment deploys the results of the optimization model to a server environment that simulates reality.From the perspective of user behavior mining,it expounds and analyzes the perceptual semantics represented by user data,and uses the perceptual semantics to evaluate the optimization effect of the optimization model.Research has proved that constructing a simulated visual competition model and adjusting it for Web visual elements can help designer better grasp the visual flow of Web pages and improve the user experience.