Research on Personalized Learning Path Planning and Career Development Linkage Based on Reinforcement Learning

Abstract

In the context of the rapid development of online education, merely providing learners with a large amount of learning resources is insufficient to meet their personalized learning needs. This paper proposes an adaptive learning path recommendation model based on reinforcement learning, exploring its potential applications in enhancing learning outcomes and connecting with career development. The knowledge tracking model (SKT) based on multi-head self-attention mechanism is applied to model the learners' constantly changing knowledge levels. The cognitive navigation algorithm is introduced to filter the candidate learning item set, enhancing the logic of the learning path and reducing the search space of the strategy function during the recommendation process. The recommendation is carried out using the reinforcement learning algorithm, incorporating the degree of change in the learners' knowledge levels into the calculation of the reward function, to evaluate the quality of the recommended items more precisely. Experimental results show that the RL4ALPR model proposed in this paper outperforms the comparison algorithms in both single-peak and multi-peak functions. The learning paths recommended based on this method are used by students, and the target LO correct rate is the highest, with a value of 0.84. These results verify the effectiveness of the RL4ALPR model in improving students' learning outcomes. It provides a feasible technical path for achieving intelligent integration of personalized learning and career development.