Evaluation of university faculty's online learning behavior using the decision-support RFM model and random forest algorithm

Online learning, as a core component of smart education, has rapidly emerged as a key driver of professional development in higher education, enabling university faculty to enhance their skills and knowledge through digital platforms. In the context of smart education, where data-driven insights, optical recognition technologies, and intelligent systems foster personalized learning and resource optimization, faculty engagement plays a pivotal role in achieving educational objectives. However, effectively analyzing and predicting faculty engagement patterns presents significant challenges due to the complexity and nonlinearity of online behaviors. This study introduces an innovative predictive framework based on the Random Forest algorithm, a powerful ensemble machine learning technique, to evaluate and forecast faculty engagement in smart educational environments. The framework integrates the Recency-Frequency-Monetary (RFM) model, adapted to smart education contexts with visual behavior analysis and optical image data, and employs feature importance analysis within Random Forest to identify key predictors of engagement. Using a dataset of over 1,200 faculty members spanning 12 months, the proposed model achieved a high prediction accuracy (R² = 0.82) and a low RMSE (0.167), demonstrating its computational effectiveness. The analysis revealed that session duration, module completion rates, and interaction patterns derived from optical recognition systems were the most significant predictors of engagement. These findings highlight the transformative potential of machine learning-based predictive frameworks, particularly Random Forest, in enabling targeted interventions and optimizing resource allocation in smart education systems leveraging optical technologies. This study contributes to the advancement of predictive learning analytics by leveraging Random Forest as a core computational tool in intelligent education ecosystems.