An Ai-Enabled Personalized Learning and Behavioral Analytics Framework for Integrating Educational Outcomes, Nutritional Awareness, And Psychological Well-Being

Abstract

Multidimensional student information can be integrated for personalized educational support, using artificial intelligence (AI) and behavioural analytics. In this research, an AI-based personalized learning framework with behaviour and psychological well-being is created. A quantitative secondary data analytical design was used, with 1,000 records of students with demographic, learning, behavioural, lifestyle, psychological, and educational variables. Descriptive statistics, group comparisons, correlation analysis, unsupervised clustering, and machine-learning approaches were used to describe the student profiles and predict the educational performance. Previous GPA showed the highest positive correlation with examination score and study hours, and motivation had positive correlation with academic performance. Examination outcomes were negatively related to stress and examination anxiety. Significant differences in examination scores existed between the various study environments and between the various diets but not between the various learning styles. Based on AI, two learner profiles were identified that differed mainly in the dimensions of motivation, examination anxiety and academic achievement. Linear regression had the best score in predicting the examination scores with R2 of 0.891, MAE of 3.20 and RMSE of 4.13 in the tested prediction models. The class imbalance in the dropout-risk classification poses a challenge for interpreting the overall classification accuracy. The results show that multidimensional profiling of learners and prediction of their educational outcomes may be achieved by combining academic, behavioural, lifestyle and psychological data with AI-based data analysis. It remains to be longitudinally and externally validated for use in educational decision-support systems.