Mining Text from Student-System Interactions to Recommend Blogs and Papers

Web-based Learning Environments (WBLE) have been used as an important space to provide different kind of interactions between students and teachers, as well as between students and software agents. However, such interactions generate thousands of data and as a consequence all the users involved are commonly overloaded of information. This fact causes a negative influence on the quality of interaction and also brings learning problems. For this reason, there is a great demand for tools to improve the dynamic of the interactions, helping the students to better explore learning materials. This work proposes an approach in which a recommending system is integrated with a WBLE for helping students in the knowledge building process by retrieving relevant information and providing recommendations of blogs and papers to them. This paper describes systems architecture and functionality, and also a preliminary evaluation in the context of teaching activities. The results of this case study with students taking a course show the effectiveness of the proposed system.