Adaptive Chinese Language Teaching and Evaluation via Big Data and Mobile Interaction

Abstract

With mobile devices becoming the mainstream platform for learning Chinese, how to utilize the big data generated by these platforms to achieve precise and personalized teaching has become a central issue. This study aims to construct a framework for Chinese language learning behavior analysis and management based on big data and interactive mobile devices. Existing research often focuses on static, unidimensional behavior analysis, with insufficient intelligence in adaptive evaluation. To address these limitations, this paper first develops a multidimensional learning behavior trajectory model that deconstructs the learning process from the cognitive, behavioral, and emotional/metacognitive dimensions. It employs lag sequence analysis (LSA) to deeply explore the temporal patterns and learning strategies embedded in the behavioral data, enabling a dynamic and detailed representation of learners’ states. Building on this, the paper proposes an adaptive teaching evaluation model, which uses the behavior trajectory features as input and innovatively applies a particle swarm optimization – backpropagation neural network (PSO-BP) hybrid neural network algorithm to solve the problem of mapping complex nonlinear behavior features to optimal teaching decisions. The model ultimately provides a personalized teaching solution that includes knowledge state diagnosis, learning path recommendations, weak point warnings, and content difficulty adjustments. This study offers a comprehensive approach-from theoretical models to technical implementation- to address the “one-size-fits-all” dilemma in mobile Chinese learning and lays a solid foundation for developing the next generation of intelligent adaptive Chinese teaching systems.