Data Analysis and Prediction of Students' Academic Performance Using Machine Learning and Deep Learning Methods

The article is devoted to the problem of predicting students' academic performance based on the analysis of educational data using machine learning and deep learning methods. The authors consider such key sources of information used in modern research as assessments and tests, attendance data, activity in learning management systems, demographic and behavioral characteristics of students, as well as the results of psychometric analysis of tests. Particular attention is paid to time series analysis, which allows identifying turning points in student learning and predicting long-term trends. The review conducted in the article demonstrates that each type of data has its own predictive potential and requires the use of specialized processing algorithms. Traditional indicators provide high reliability but are limited in terms of time, while dynamic LMS (Learning Management System) data and time series allow for early risk detection. Hybrid approaches that integrate multiple data sources and analysis methods are the most promising, as they enable the construction of adaptive learning support systems