Key points are not available for this paper at this time.
The inclusion of e-learning platforms in the traditional education allows obtaining a lot of data about the students' behavior. Studying these data using data mining and learning analytics techniques allows detecting different behavior patterns and predicting future results. One of the big challenges in education is being able to predict the students' failure before it occurs, and avoid it. Being aware of this challenge, and inside the learning analytics context, this paper describes a detection risk algorithm that tries to detect the students who are at risk of failing the course based on their interaction with an e-learning platform. The algorithm is based on time series formed by the daily students' interactions with the platform. More concretely, the trend component of the time series is used as a predictor for the students' results. We apply our algorithm to one course of the 2nd-year of the Telecommunication Engineering Degree at the University of Vigo, obtaining some encouraging results.
Nespereira et al. (Wed,) studied this question.