Background: Advancements in machine learning (ML) have revolutionized the analysis of complex educational data. Algorithms such as k-Nearest Neighbors (KNN) and Support Vector Machines (SVM) enable early identification of students at risk of academic failure, facilitating timely interventions. This study reviews the advantages, limitations, opportunities, and challenges associated with ML in predicting student success.
Materials and Methods: In this narrative review, a literature search was conducted in both Persian and English across online databases, including Scopus, Web of Science, IEEE Xplore, SID, CIVILICA, and Google Scholar, covering studies up to March 2025. Two researchers independently screened and selected relevant articles and extracted key data.
Results: Analysis of the literature reveals that machine learning algorithms, such as K-Nearest Neighbors (KNN), Support Vector Machines (SVM), and deep learning, offer substantial improvements in predicting academic success and tracking students’ academic progression. Their high predictive accuracy supports the timely identification of at-risk students, facilitates data-driven educational decisions, and enables scalable, adaptive interventions. Nonetheless, the full impact of these technologies depends on the availability of high-quality data and is tempered by challenges such as algorithmic bias, limited interpretability of complex models, technical and resource constraints, and ethical concerns regarding privacy and fairness. Looking forward, integrating explainable models, enhancing adaptability to diverse educational contexts, and leveraging personalized, technology-driven learning environments are identified as pivotal to maximizing the educational benefits of machine learning.
Conclusion: Machine learning significantly enhances education by accurately predicting and early identifying at-risk students. However, unlocking its full potential requires addressing challenges related to data quality, privacy, bias, transparency, and infrastructure. Future progress depends on developing flexible, fair, and culturally sensitive models to ensure the ethical and effective application of ML across diverse educational environments.