Abstract
The high rate of student dropout, especially due to failure in the first months of training, remains an urgent problem of modern education, posing both economic and personnel risks. This fact served as the basis for considering in this study the possibility of creating a system of predicting the success of students based on mathematical statistical methods. The purpose of the work is empirical confirmation of the relationship between the activity of attending lectures and the success of students as a basis for the development of a digital assistant. With the help of correlation-regression analysis, a statistically significant positive relationship between the considered indicators was established. The obtained results allow us to conclude that attendance is one of the main predictors of good student performance. The developed model allows predicting the risk of failure at an early stage, timely informing students and the university administration, which contributes to improving the quality of the educational process and reducing the probability of student dropout.