Bridging Technology and Psychology: AI-Driven Analysis of Student’s Class Lecture Activity for Improved Learning Outcomes

Abstract

Students’ emotional state and attention significantly impact how they handle stress and interact

with their studies. These factors are crucial in defining their learning objectives and general personal growth,

influencing their academic achievement. Because Bangladesh has distinct educational and mental health

difficulties, it is important to comprehend these dynamics. Time series analysis is a useful technique for

tracking lecture activities in class and how they affect student participation since it provides insightful

information about behavior patterns across time. This investigation aims to address these problems by using

MotionWatch 8 and a comprehensive questionnaire to analyze class lecture activities. This study employs

ensemble methodologies, deep Learning algorithms, and a diverse range of machine learning models to

assess and predict student behavior. A hybrid model is one of the techniques that produced the most stunning

results, proving how well it could capture complex patterns in time series data. To assess the robustness of the

algorithms, the study also looked at how different datasets performed. Ultimately, the models’ interpretability

was improved, and their decision-making processes were given a more profound understanding by utilizing

explainable AI methods, including SHAP, LIME, and permutation importance. The effort establishes new

standards for improving student engagement and well-being through data-driven insights, sophisticated

models, and explainable AI.