Using Process Mining algorithms to identify patterns of student behavior

This work is devoted to studying the possibilities of using process analytics methods (Process Mining) to analyze student activity based on digital traces that students leave in learning management systems (LMS). This work examines the specifics of process mining algorithms that can be used to analyze educational processes, namely, heuristic and inductive algorithms are considered as the most effective for building models and suitable for use for the purpose of analyzing educational data. The work involved creating a way to use process analytics algorithms to identify clusters of students with similar behavior patterns. The development of a process analysis algorithm was carried out on the basis of the event log of the distance learning system of Kostroma State University. As a result of the work, models of student behavior were built and visualized, including the identification and clustering of students with similar behavior, the construction of heuristic networks, Petri nets, a direct sequence graph, a BPMN model and a decision tree. An analysis of the resulting models was carried out, which showed that the developed method makes it possible to study the behavioral patterns of students. The proposed method of using intellectual analysis of educational processes can be used to solve issues of increasing the productivity of the educational process, early detection of problems, especially in the context of changing student behavior in the system, as well as the development and optimization of educational programs. In addition, the limitations of this system have been identified, which may hinder its implementation and application in the educational environment of universities.