A study of learner capability portrait building and customized recommendation algorithms via behavioral data and optical data

The rapid development of digital learning environments has generated vast amounts of behavioral data, yet existing learner modeling approaches often fail to capture the multidimensional nature of learning capabilities. This study addresses the critical need for advanced analytical methods to construct comprehensive learner portraits and deliver personalized recommendations. The research aims to develop a novel framework that integrates behavioral data analytics with adaptive recommendation algorithms, specifically designed for educational contexts. The methodology combines quantitative analysis of interaction logs with qualitative behavioral pattern recognition to build dynamic learner profiles. Optical methods, including facial expression analysis and eye tracking, enhance behavioral data capture for more precise learner modeling. A hybrid recommendation algorithm is proposed, incorporating competency-based filtering and real-time adaptation mechanisms. Experimental validation involves data collected from 320 participants across diverse learning scenarios, with system performance evaluated through both accuracy metrics and user experience assessments. Results demonstrate significant improvements in recommendation precision, with the proposed algorithm achieving 23% higher accuracy compared to baseline methods. The learner portrait system successfully identifies distinct capability patterns across cognitive, metacognitive, and affective dimensions. User studies reveal strong positive correlations between system recommendations and actual learning needs, as evidenced by 82% satisfaction rates in practical deployment scenarios. This research contributes to the field of educational technology by providing a scalable framework for behavior-driven learner modeling. The findings offer practical insights for developing next-generation adaptive learning systems that can effectively translate behavioral data into personalized educational support.