Collaborative recommender system using distributed rule mining for improving web-based adaptive courses

Nowadays, the application of data mining techniques in e-learning and web based adaptive educational system is increasing exponentially. The discovered useful information can be used directly by the teacher or the author of the course to improve the instructional/learning performance. This can be an arduous task and therefore educational recommender systems are used in order to help the teacher in this task. In this paper we are going to describe a recommender system oriented to suggesting the most appropriate modifications to the teacher in order to improve the effectiveness of the course. We propose to use a cyclical methodology to develop and carry out the maintenance of web-based courses in which we have added a specific data mining step. We have developed a distributed rule mining system in order to discover information in form of IF-THEN recommendation rules about the web courses. We have used an iterative and interactive association rule algorithm without parameters and with a weight-based evaluation measure of the rule interest. And we have used a collaborative recommender system to share and score the obtained recommendation rules in one specific course between teachers of other similar courses and some experts in education. Finally, we have carried out several experiments with real students in order to determine the effectiveness of the proposed system and the utility of the recommended rules.