Enhancing adult lifelong learning through data fusion algorithms and quality assurance

Abstract

This study presents a Data Fusion + Random Forest (RF) framework to predict adult learners’ educational outcomes, such as dropout risk and course completion. Adult lifelong learning is crucial for personal and professional growth, but challenges like high dropout rates and low engagement persist. Existing models often fail to account for the complexity of adult learners’ behaviours, leading to ineffective predictions. The objective of this work is to improve prediction accuracy by combining demographic, behavioural, and performance-related features into unified learner profiles, using the Adult Learning Dataset. The proposed framework employs a Random Forest classifier, providing enhanced predictive power. Key findings show that the proposed method achieves an accuracy of 97.88%, precision of 99.87%, and recall of 95.77%, outperforming existing models, which report accuracy around 80% and F1-scores of 0.83. These numerical improvements demonstrate the efficacy of the proposed framework in handling diverse learner data. The model is further optimised using hyperparameter tuning, contributing to its reliability and robustness. The results emphasise the potential of this framework to support data-driven interventions in adult education, improving retention and engagement. Its practical applications could significantly enhance the prediction of learner outcomes, leading to more tailored and effective educational strategies.