Warnings and Recommendation System for an E-Learning Platform

Abstract

A warning messages and recommendation system for an E-Learning system is proposed, the goal is to identify which students are likely to have a poor academic performance, and give them timely feedback by showing alerts and recommended material. The proposed system uses a set of profiles previously identified by a student profiling model, using socio-economic (age and gender) and web navigation data on the system (number of accesses to resources, percentage of accesses in class, average absence time and average session length). Each profile is analyzed and a warning message is assigned to each one; also, the sequences of consultations performed by students with a high academic performance are recognized and used to choose which resources are recommended. Based on the sequence performed by a student in a current session, the platform may recommend access specific resources.