Um sistema de recomendação para professores e coordenadores de curso utilizando predição de reprovação na modalidade de educação a distância

Abstract

This text proposes an Educational Recommendation System (ERS) model for teachers and course coordinators based on prediction of apprentices failure in e-learning courses. Considering the role of education in shaping the individual, it is important to keep the apprentice in the course until it's completion, seeking to provide a good GPA. E-learning has expanded the options of people seeking a qualification, but for it's success the numbers, at the end of the course, need to show a positive performance. Educational Recommender Systems have been developed to explore the profiles of learners in order to detect their problems and to make them a suggestion, which can be a learning object, a literature review, or a collaborative study with other learners. The aim of this work is to propose M-SRECP, an educational recommender system model whose target audience is course coordinators and teachers, who receives statistics of possibility of failure of apprentices in a discipline, from a prediction system, and makes recommendations to this target audience, so they can make pedagogical actions aiming to reduce the number of failed apprentices. Techniques such as user profiles classification that guide the delivery of the recommendation, context awareness and Customer Relationship Manager were used in this work, in order to provide a differential on the recommendation to the target audience (teacher and course coordinator ), putting them in the main focus of the process (CRM), over a discipline (context awareness). The actions of the target audience aims at reducing the failed apprentices in e-learning mode disciplines, offering them an opportunity to improve their GPA, reduction of the time enrolled in the course and thus accelerating their certification. A prototype of the model was developed for an experiment at the Federal Institute of Education, Science and Technology of Amazonas, in the Open University of Brazil program, in the Philosophy of Education course, in discipline Brazilian Language of Signs in 2015/2. It was collected through a questionnaire, 30 teachers profiles, which allowed the classification of teacher's profile using Decision Tree with RapidMiner. The discipline Libras teacher's profile was also collected, however it has been subjected to classification to determine whether, besides him, there would be the need for the coordination's course of Philosophy of Education also receive weekly recommendation of apprentices with failure risk in the discipline. The prototype was also presented to a set of 12 teachers so that they could make an evaluation of perceived ease of use and utility perception through the Technology Acceptance Model (TAM). It was concluded that M-SRECP is a computational tool that can help teachers and course coordinators to rescue apprentices before the failure to occur, helping to improve the learner's performance in a discipline still in progress.