The work updates a strategic approach to maintaining the quality of student performance. In the article "Big Data" is presented as an approach to organizing the data obtained during the educational process. Mining analysis as a Big Data analysis tool. The authors also considered approaches to improving the effectiveness of learning, which allow predicting student performance. Blended learning, as an integration into the educational process, the best opportunities for online and distance learning, provided for the subsequent collection of Big Data for analysis. The application of data mining methods for the development of models for predicting the behavior and academic performance of students is considered. The authors propose student modeling as a key concept in data mining in education, which refers to the qualitative representation of student behavior. Algorithms for data mining are described, classification and clustering are presented as the most common tasks. The CART decision tree algorithm is presented as one of the effective data mining methods needed to predict student performance based on online activity that is stored in the Moodle LMS log file. Finally, it was concluded that the personal contact of an experienced teacher with a student in the educational process can not be replaced by the tools provided by Moodle and similar systems, but due to the growing threats caused by the pandemic, the relevance of distance education is becoming increasingly important.

Keywords: student performance quality, big data, data mining, student modeling, CART decision tree algorithm, Moodle LMS, distance education