USING THE MACHINE LEARNING METHOD TO ANALYZE THE DYNAMICS OF THE PERFORMANCE OF PRACTICAL TASKS IN COMPUTER SCIENCES

Abstract

Covid-19 quarantine restrictions and the war in Ukraine have actualized distance learning more than ever. In these conditions, management of the learning process, recording and analysis of learning results of large groups of education seekers became very important. Management of any process involves feedback. In this case, the analysis of relevant statistical distributions of training results can become a relevant perspective direction of providing feedback. In this work, the methods of machine learning, namely: the regression tree and the classification tree, were selected as a scientific methodology that ensures the above-mentioned analysis. The purpose of the work is to study the dynamics of learning processes of large groups of education seekers using machine learning methods, which is realized by analyzing the results of measuring changes in the speed of performing practical tasks in informatics. The scientific novelty of this work consists of an attempt to apply machine learning methods to the analysis of the results of practical educational activities. In the course of research, more than five thousand relevant measurements were taken. In addition a program was written for the analysis of these data in the Python language using the Scikit-Learn library. This program is presented in the work. With the help of machine learning methods, an analysis of the results of measuring the speed of computer science tasks was carried out. The resulting graphs have a smooth shape with a slight bend, without extremes. Significant linear fragments are fixed on the graphs. The paper compares the results of the data analysis with the results of the analysis of the same data carried out earlier using the method of moments. There is a coincidence of related dependencies of the dynamics of changes in the speed of educational actions, obtained by the regression tree method and the method of moments.