A Data-Driven Approach to Personalized Learning And Course Enhancement in Education Technology

This research focuses on enhancing online learning experiences by analysing student engagement through behavioural and performance data. Using a dataset comprising quiz scores, assignment completion rates, feedback, and visual cues (like confusion, boredom, and gaze aversion), we classify engagement levels using machine learning algorithms such as Decision Tree, Random Forest, SVM, k-NN, and Logistic Regression. Additionally, image-based features are incorporated to detect emotional states, enabling accurate classification of engagement into Low, Medium, and High levels. Personalized suggestions are generated for each student using decision trees based on their learning behaviour and engagement metrics. Furthermore, collective course-level analysis reveals trends and informs recommendations for improving curriculum design.