Virtual Classrooms as Data Sources for Prediction Tools

Abstract

Nowadays, on-line campus are very important in the learning process of students, since they can access to teacher's resources easily. Moreover, on-line campus provide useful tools for building evaluation processes by teachers. Under this point of view, knowing the strengths and weakness of a student before his evaluation allows to plan better his learning schedule when we analyze the history of the student and his colleagues, in the same tasks. In addition, the instructor can predict the difficulty level of the proposed exercises for determined students, allowing him to adjust better the evaluation tasks. Predicting the student performance can be obtained from Machine Learning tools, specifically Collaborative Filtering techniques based on Recommender Systems. In this paper, we explain how we are using and including these techniques in a Moodle environment, in order to provide several useful resources for both students and teachers. In this context, virtual classrooms provide useful data for predicting purposes.