ODKT-IAC: Integrating Continual and Adaptive Learning for Personalized Path Recommendations

In today’s educational landscape, the diverse needs and preferences of learners make Personalized Learning Path Recommendation Systems (PLP-RS) essential for enhancing learning experiences and outcomes. This paper introduces our solution, ODKT-IAC, that integrates Continual Learning (CL) with Adaptive Learning (AL) to ensure the system adapts to continuously changing learner needs. Initially, AL techniques are employed to analyze learning data, including goals, and prior history, to create detailed learning profiles by identifying strengths and weaknesses. Subsequently, CL algorithms, such as online learning in Knowledge Tracing (KT) and incremental Reinforcement Learning (RL), are used to update the model with new learning data, behaviors, and scores. Online learning allows real-time tracking and updating of learners’ knowledge states, while incremental RL enables the algorithm to learn from new data and adjust accordingly, maintaining learner motivation and interest. We evaluate our approach using the Online Learning Activity dataset from Junyi Academy, which includes comprehensive learning logs from 247,548 students over the course of a year. Experimental results demonstrate that our approach enhances learning efficiency in dynamic environments and exhibits superior adaptability and effectiveness compared to traditional methods. Our solution paves the way for intelligent learning recommendation systems (RS), enhancing educational quality, and supporting learners in achieving their academic and career goals.