Early Multi-criteria Detection of Students at Risk of Failure

In this paper, we present a new fuzzy methodology for early students’ failure detection. High school background, subjects studied in the university and activities in learning management systems were determined to be the factors influencing students’ performance. After selection of the impact factors of students’ assessment, we convert linguistic evaluations into fuzzy numbers and employ multi-criteria methods for educational data processing. In two practical examples, the aggregate students’ scores were calculated by using fuzzy multi-criteria algorithms. The obtained students’ ranking helps instructors during the semester to detect students who will drop out the course and to plan additional learning activities for these students. In the future, we plan on analysing students’ data from different university’s courses and majors and mining several academic years in order to create a reliable assessment index for early prediction of students’ failure.