The paper presents a hybrid AI framework for academic advising that combines an ensemble-based grade prediction model with a rule‑based expert system. Students from the University of Birjand were clustered using Gaussian Mixture Models, and a Stacking Ensemble of Random Forest, Gradient Boosting, and MLP was trained per cluster, achieving an aggregated RMSE of 2.35. The expert system then uses these predictions alongside institutional regulations to give real‑time feedback such as GPA forecasts, probation warnings, and course recommendations.
By Hamid Saadatfar, Rohollah Hedayati-Nasab, AmirHossein Eshghi, Arash Hajihashemi
arXiv:2607. 14124v1 Announce Type: cross Abstract: The increasing availability of large-scale educational datasets has expanded the use of quantitative methods for investigating school performance.
By Anderson L. de Paula, Pedro C. dos Santos, Renato A. Krohling
arXiv:2608. 13409v1 Announce Type: new Abstract: Existing predictive models in learning analytics often treat student academic history as a simple sequence, overlooking the concurrent nature of courses taken within a semester.
By Paul Savala
arXiv:2607. 26063v1 Announce Type: cross Abstract: Personalised learning systems often assume that mathematical ability is combined of discrete abilities, acquired sequentially and dependent upon first acquiring foundational abilities, and students often report different strengths.
By Benjamin Mawdsley, Tom Quilter, Richard Turner, Sarah Jackson, Paul Edwards
arXiv:2608.29110v1 Announce Type: new
Abstract: The United States has allocated approximately $65 billion through the Infrastructure Investment and Jobs Act for broadband expansion, yet evidence-base...
By Xiao Han
The paper presents a data‑driven study of male domestic violence (MDV) in Bangladesh, using exploratory data analysis to uncover patterns such as verbal abuse prevalence and the influence of financial dependency. It evaluates 10 traditional ML models, 3 deep learning models, and 2 ensemble models, ultimately proposing a stacking ensemble with ANN and CatBoost base classifiers and Logistic Regression meta‑model that achieves 95% accuracy and 99.29% AUC. Explainable AI techniques (SHAP, LIME) and statistical validation confirm the model’s superior performance and highlight key features driving predictions.
By Md Abrar Jahin, Saleh Akram Naife, Fatema Tuj Johora Lima, M. F. Mridha, Md. Jakir Hossen