An Intelligent Feedback Mechanism for Mobile English Learning Based on Learning Behavior Analytics

Abstract

The rapid proliferation of mobile learning technologies has reshaped English learning contexts, giving rise to fragmented learning patterns and increasing demands for personalization. However, feedback mechanisms in current mobile English learning systems remain overly focused on knowledge error correction, with limited consideration of learners’ cognitive processes, thereby constraining learning effectiveness. The deep integration of learning behavior analytics with intelligent feedback is regarded as a promising pathway to address this limitation. This study aims to address a central research question: how can fine-grained learning behavior analytics be leveraged to integrate observable behavioral data with latent metacognitive states in order to construct a pedagogically adaptive and personalized intelligent feedback mechanism for mobile English learning? To this end, a multi-task deep knowledge tracing (DKT) approach incorporating metacognitive assessment was proposed. On this basis, a closed-loop framework integrating data perception, joint modeling, and multidimensional feedback was constructed, with optimization strategies tailored to mobile learning scenarios. Experimental results based on real-world mobile learning data demonstrated that the proposed approach significantly outperformed conventional DKT models in both knowledge tracing accuracy and metacognitive state recognition accuracy. Moreover, the resulting intelligent feedback mechanism effectively enhanced learners’ English learning performance and metacognitive abilities. This study extends existing language learning theory through the integration of metacognitive modeling with knowledge tracing, introduces a discipline-adaptive multi-task intelligent feedback modeling paradigm, and provides a practical pathway for the intelligent enhancement of mobile English learning systems.