Aug 2026· Herald of Dagestan State Technical University. Technical Sciences· Vol 53, pp. 23-35· 0 citations· 10 references
TL;DR
The proposed approach enables a transition from reactive, exam-focused preparation for state final certification to proactive management of the educational trajectory, resulting in a 10-point increase in the average Unified State Exam (USE) score.
Abstract
Objective
. The study aims to identify the relationship between the quality of education in general educational institutions and students' cognitive and personality traits, as well as parameters of the educational environment, with the goal of developing a tool for forecasting and adjusting subject-specific learning outcomes.
Method
. Based on surveys of participants in the educational process at the online school «Dom Znaniy» (House of Knowledge), a database comprising over 1.3 million records across 315 parameters was compiled. After filtering, 218 significant indicators were retained. An original iterative method utilizing neural networks was applied to restore missing data, and a novel approach based on decision trees was proposed for vectorizing categorical features. An artificial neural network was trained on data from 2022–2023 to forecast subject learning outcomes.
Results
. A neural network was developed capable of predicting students' subject-specific results for the upcoming academic year with an error margin of less than four percent. All 218 parameters were ranked by their influence on academic performance, and the quantitative relationship of each parameter to the final grade was established. A user interface was created enabling parents to simulate changes in educational environment parameters and a child's personal characteristics to achieve target educational outcomes.
Conclusion
. The proposed approach enables a transition from reactive, exam-focused preparation for state final certification to proactive management of the educational trajectory, resulting in a 10-point increase in the average Unified State Exam (USE) score. Future development of the project includes automating parameter collection using artificial intelligence technologies.
Seven machine learning methods were applied to classify and predict student academic performance: Multinomial Naïve Bayes, Support Vector Machines, Decision Trees, Logistic Regression, Random Forest, Gradient Boosting and k-Nearest Neighbors.
The Explainable Learning Analytics Framework (ELAF) as an integrated eight-layer conceptual architecture is suggested as an integrated eight-layer conceptual architecture to cover this persistent gap between the prediction accuracy of algorithms and the use of the algorithms in the educational domain.
Pratik Mishra, Nishesh Bishwas, Anjana Neupane· Apex Journal of Business and...· 0 citations
Predicting how well a student will perform before, rather than after, a semester goes wrong is one of the more practical promises that machine learning has brought to higher education. This study reports a case study conducted at a public university, in which academic records, attendance logs, and a small set of socio-...
Habibrahman Habibi, Abdul Wajid Fazil, Kefayatullah Khairkhah· Frontiers in Sustainable Sci...· 0 citations
Predicting student academic performance was important for identifying students at risk of poor achievement and supporting timely educational interventions. This study developed and evaluated an XGBoost-based machine learning model for predicting student performance using academic, behavioural, and socioeconomic factors...
A. Yahaya, Muktar Danlami, Zawali S. Paki et al.· British journal of computer,...· 0 citations