Intelligent Decision Support for Assistant Selection for Checking Works with Free Response in Massive Open Online Courses with the Use of Text Mining

In massive open online courses (MOOCs), a teacher’s assistant role is very significant because the most important functions of tutors are delegated to assistants. Considering a huge number of students on every launch of the MOOC, the tutor is not able to serve it alone. The studies associated with the use of teacher’s assistants in online courses are focused on different aspects. However, the former do not address the question connected with the consideration of assistants’ ability to form response to performed works with free response (WWFRs) for the students properly, in detail and close to tutor’s feedback. The response obtained by the student will determine whether the required skills will be formed at him. This problem is particularly serious for engineering education with the presence of a large number of WWFRs: practical works, labs and other works. To cope with this issue, a method of intelligent decision support for assistant selection for checking WWFRs is suggested in the paper. The method contains the steps: the preliminary selection of assistants on eliminating criteria; organization of performance of assessment of works, earlier evaluated by the tutor, by assistants-candidates; definition of the quality of evaluation conducted by assistants-candidates by identifying the degree of similarity of their and tutor’s assessment results, employing text-mining. The proposed method was implemented in software using Moodle based learning management system and author’s application and tested on data on course “Information resource management” at university.