No abstract

No abstract

Objectives The purpose of this study was to explore and define the factors that constitute self-directed learning in smart learning from a learning analytics perspective and to validate the appropriateness of the type and characteristics of the learning data. Methods To ensure the validity of the learning analytics data, we used an expert validation method. For this purpose, a total of seven experts with over 8 years of experience in e-learning, educational research, and educational practice participated. We explained the target service of this study and the purpose and tools of the validation test to each expert and conducted the test twice via email. Results We defined the self-directed learning data that can be collected and analyzed among various data generated throughout the learning process in smart learning. Specifically, we classified them into three factors of self-directed learning: 1) meta-cognition (planning, monitoring, self-evaluation, and reflection), 2) learning strategies (performance, review, problem-solving strategies, elaboration, and exploratory learning strategies), and 3) learning behaviors (time management, learning performance, immersion, concentration, problem-solving habits, and social learning participation) and a total of 41 detailed learning analysis data were derived. Conclusions This study provides meaningful directions for smart learning system to support learners' self-directed learning under the learning analytics. Also, it is expected to help lay the foundations for future AI-based learning system.