Multi-criteria fuzzy ordinal peer assessment for MOOCs

Due to the high number of students enrolled and the relatively small number of available tutors, the assessment of complex assignments is deemed as one of the most critical tasks in Massive Open On-line Courses (MOOCs). Peer assessment is becoming an increasingly popular tool to face this problem and many approaches have been proposed so far to make its outcomes more reliable. A promising approach is FOPA (Fuzzy Ordinal Peer Assessment) that adopts and integrates models coming from Fuzzy Set Theory and Group Decision Making. In this paper we propose a FOPA extension supporting multi-criteria assessment based on rubrics. Students are asked to rank a small number of peer submissions against specified criteria, provided rankings are then transformed in fuzzy preference relations, expanded to obtain missing values and aggregated to establish a global ranking between students' works with respect to each criterion and globally. The absolute grades of all submissions are then calculated.