Abstract
One of the main goals of any educational institution is high quality education, taking into account the shortcomings of the education system. Currently, in the field of education, one of the urgent problems is the identification of signs that affect student performance. Machine learning algorithms are used to determine these features. Machine learning algorithms are being implemented in various spheres of society today. Giants from all over the world, such as Amazon, Google, Apple, identify customer-related problems using machine learning algorithms. In addition, a number of machine learning methods in economics, banking and medicine, manufacturing and industry can improve the work of specialists. Aiming at better performance and better understanding of the obtained results, binding models had been applied at the data selection and integration stages. The article discusses the collection of databases, learning algorithms, analysis of algorithms and the definition of features. In the research work, a database is created that includes the physiological and psychological characteristics of students. The program is executed in the Python programming environment. Predictions are based on databases with machine learning algorithms. Linear regression methods, the Support Vector Machine method and Random Forest are used, and work is devoted to determine the best algorithm. Algorithms are evaluated using regression evaluation metrics. The results of the study provide the best algorithm and the necessary marks for individual academic performance.