The Application of Computer Technology in User Behavior Analysis and Optimization of Intelligent Education Systems

Current behavior modeling often ignores the interaction of variables such as learning interest, psychological state, and skill mastery, and lacks accurate capture of students' dynamic behavior. This paper explores the application of computer technology in user behavior analysis and optimization in intelligent education systems, focusing on the modeling of learner behavior. Since learners' behavior is affected by many factors, such as learning interest, psychological state, and skill mastery, their behavior is highly variable and complex. First, the study revolves around several key topics of user behavior analysis, including traffic analysis, user portrait analysis, time dimension trend analysis, and user interest analysis. Based on these analyses, this paper further proposes a model for real-time evaluation of learner status, which combines learners' test scores and video viewing behavior to comprehensively reflect learners' learning status. Next, this paper introduces a dynamic routing algorithm that uses capsule networks to capture learners' interest changes during the learning process. With this method, learners' interest fluctuations at different learning stages can be tracked more accurately, and their behavioral changes can be modeled. Finally, through comparative experiments, this paper verifies the learner status judgment method based on video viewing behavior and test scores, which can more accurately reflect learners' cognitive status and learning progress compared to traditional models that only rely on single indicators of scores or behaviors.