Abstract
This research proposes a hybrid predictive model designed to identify at-risk students within a gamified education environment accurately. By integrating logistic regression, decision trees, and random forests, we construct a robust ensemble model that leverages the strengths of each algorithm for precise risk assessment. The model analyzes key indicators such as academic performance, participation levels, and task completion rates using data derived from a gamified learning platform. Our approach demonstrates the effectiveness of machine learning in addressing challenges like student disengagement and dropout. The hybrid model outperforms individual classifiers, enabling earlier and more reliable detection of students who may require timely academic interventions. The method is as follows:•Combines logistic regression, decision trees, and random forests•Utilizes gamified education data for at-risk student prediction•Provides educators with a tool for early intervention in student supportThe computational approach converts raw educational data into actionable insights, enabling educators to deliver timely and targeted interventions. Leveraging behavioral data from game-based learning platforms, the project develops a practical student monitoring system powered by machine learning ensembles. This system identifies at-risk students earlier than traditional assessments, allowing for more effective and efficient use of educational resources.