Stress Detection and Decision Making Using Natural Language Processing and Fuzzy Sets to Prevent Burnout in Transformed Online Learning Platforms
Аннотация
The fast shift to online learning platforms has made education more adaptable and accessible but has also presented such issues as the stress levels of learners, emotional exhaustion, and academic burnout. Such issues as the extended screen time, individual requirements of learning, lack of socialization, and the pressure of the performance also contribute to the increase of psychological stress among the learners. Early identification of stress and a timely response to that is hence important in maintaining the engagement and academic performance. The given framework examines textual data, which is created by learners during discussion forums, chats, feedback, and reflective submissions and applies techniques of Natural Language Processing to get linguistic, sentimental, emotional, and stress-related features. Natural language is naturally ambiguous and thus, fuzzy logic is employed to deal with ambiguity in interpretation of emotions. Depending on the levels of stress identified, the system allows students to implement adaptive interventions: learners with low levels of stress proceed with normal learning, those with moderate stress levels get motivational support and workload related adjustments, and learners with high risks activate immediate interventions. Experimental evidence indicates that the hybrid NLP-fuzzy model is also superior in terms of accuracy, precision and in recall, and thus it can be used in intelligent online learning systems.