Data mining of students' behavior in E-learning system

Abstract

The article deals with educational data mining techniques aimed at increasing effectiveness of E-learning process as well as the idea of adaptive feedback, individual assessment and more personalized attention to student's profile due to dynamic monitoring and tracking of students' behavior in the E-learning system. The following techniques are identified: cluster analysis to determine the most popular time threshold for the task per session; analysis and visualization of data to highlight the main options that contribute to the effective completion of courses, and the most popular educational resources; V-fold cross-checking with the use of statistical processing aimed at students by their main indicators of activity to determine the correlation between high percentage of activity and academic performance. The proposed educational data mining techniques allow to assess student's behavior in the E-learning system for understanding student's interest in studying the learning materials and assessing the quality of educational content.