Machine Learning-Based Early Detection of Learning Gaps in K–12 Students
Abstract
Early identification of learning gaps is essential for providing timely educational support and preventing persistent academic difficulties among K–12 students. Traditional assessment approaches often identify learning problems after considerable academic decline, limiting opportunities for early intervention. This study developed a machine learning-based predictive framework for detecting learning gaps by integrating academic, behavioural, and engagement-related indicators. A student learning dataset was analysed through data preprocessing, feature engineering, and predictive modelling approaches. Multiple machine learning algorithms, including Logistic Regression, Support Vector Machine, Random Forest, Extreme Gradient Boosting (XGBoost), and Artificial Neural Network, were developed and evaluated using accuracy, precision, recall, F1-score, ROC-AUC, and confusion matrix analysis. The comparative results demonstrated that ensemble learning approaches provided superior predictive performance by effectively capturing complex relationships among different learning indicators. XGBoost achieved the highest classification performance, followed by Random Forest, indicating strong capability for identifying students at risk of learning difficulties. Feature importance analysis revealed that previous academic achievement, assessment performance trends, assignment completion behaviour, and engagement patterns were major contributors to learning-gap prediction. The findings highlight the potential of artificial intelligence-driven educational analytics for transforming traditional reactive assessment systems into proactive early-warning mechanisms. The proposed framework provides educators with data-driven insights for identifying vulnerable learners and designing targeted interventions. This research contributes to the growing field of artificial intelligence in education by presenting an effective predictive approach for supporting personalised learning and evidence-based educational decision-making.
How to Cite
Nidhi Bhatia, Dr Sarika Sharma, "Machine Learning-Based Early Detection of Learning Gaps in K–12 Students", Vol. 2, Issue 12, 29-03-2025, pp. 35-48.
