Estimating learning outcomes in IoT-enabled smart classrooms using real-time data insights

This article discusses the estimation of learning outcomes for IoT-enabled smart classrooms to enhance educational effectiveness with real-time data. Based on the presented approach, this study identifies key factors influencing student performance using IoT technologies and advanced machine learning models, aiming to enhance data-driven, personalized teaching. Two different approaches, XGBoost and Light Gradient Boosting, have been used to estimate learning outcomes. In this, it was optimized using the bald eagle search (BES) optimization and coati optimization algorithms (COAs), both developed for hyperparameter tuning and to enhance predictive accuracy. First, is the valuable contribution given to this research – the incorporation of K-fold cross-validation into the model evaluation routines supported sensitivity analysis on the most influential features regarding the learning outcomes. Among four identified ensembles, the BESO-optimized XGB came out as supreme with 0.982 for Accuracy. This could be clearly explained by the contribution of various metrics to the values generated; “students’ engagement measures” are considered the major influential variable. It provides a comprehensive framework that combines data from the IoT, ML models, and optimization methods to offer useful insights for designing and implementing smart classrooms that will help maximize students’ learning outcomes.