Many of the articles on AI in education compare the performance and fairness of different models, but few specifically focus on quantitatively analyzing their explainability. To bridge this gap, we analyzed key evaluation metrics for two machine learning models—ANN and DT—with a focus on their performance and explainability in predicting student outcomes using the OULAD. The methodology involved evaluating the DT, an intrinsically explainable model, against the more complex ANN, which requires post hoc explainability techniques. The results show that, although the feature-based and structured decision-making process of the DT facilitates natural interpretability, it struggles to model complex data relationships, often leading to misclassification. In contrast, the ANN demonstrated higher accuracy and stability but lacked transparency. Crucially, the ANN showed great fidelity in result predictions when it used the LIME and SHAP methods. The results of the experiments verify that the ANN consistently outperformed the DT in prediction accuracy and stability, especially with the LIME method. However, improving the interpretability of ANN models remains a challenge for future research.

Artificial intelligence (AI) is gaining traction for its ability to extract valuable information from vast amounts of data, despite concerns about a lack of transparency in decision-making processes. The rise of explainable AI (XAI) has improved human comprehension and handling of AI systems by providing clear explanations for their decisions. This study aims to present a comprehensive overview of recent research on explainability in Learning Analytics (LA) and Educational Data Mining (EDM) from 2009 to July 2025. Through a detailed evaluation of experimental studies and adherence to the PRISMA guidelines, we initially discovered 1,656 studies, which were subsequently narrowed down to a final corpus of 212 studies for in-depth systematic analysis. Six databases were examined using a systematic keyword search: IEEE Xplore Digital Library, ACM Digital Library, Springer, Web of Science, ScienceDirect, and SCOPUS. This methodology provides a comprehensive summary of the present status of explainability studies regarding educational models and provides details about the results, methods, techniques, and efficacy of explainability uses in the realm of education. In particular, researchers have increasingly adopted post-hoc methods—especially SHAP—from 2024 to 2025, signaling a shift toward interpretable, model-agnostic tools. It also addresses the influence of metrics, models, and data types on explainability. Interestingly, only a few papers in our collection included quantified explanations of prediction models that used metrics related to explainability, such as sensitivity and stability. Nevertheless, within our corpus, several articles focused on measures related to model performance and fairness in educational settings. At the end of the review, significant findings about explainability in EDM and LA are summarized, along with a discussion of the research’s limitations and future research.