MACHINE LEARNING METHODS FOR BEHAVIORAL AND LINGUISTIC DATA ANALYSIS OF LEARNERS IN COMPUTER-BASED LEARNING SYSTEMS

The rapid digitalization of education has led to the widespread adoption of computer‑based learning systems, which continuously generate large volumes of learner interaction data. These data include behavioral interaction logs and textual responses created by students during the learning process. The present study proposes amultimodalmachine learning framework for analyzing behavioral and linguistic learner data in order to predict academic performance and evaluate cognitive engagement in digital learning environments. Behavioral indicators are extracted from interaction logs, while linguistic features are obtained using natural language processing techniques such as TF‑IDF vectorization and contextual embeddings. Several supervised learning algorithms, including Random Forest, Support Vector Machines, Gradient Boosting, and neural network models, are evaluated and compared. Experimental results demonstrate that hybrid multimodal models that combine behavioral and linguistic featuressignificantly outperformmodels based on a single data modality. The findings highlight the importance of integrating behavioral analytics and language mining for the development of adaptive learning env