Predição de desempenho de estudantes em ambientes virtuais de aprendizagem

Abstract

In this work, we investigate the using of Educational Data Mining (EDM) techniques based on Principal Components Analysis (PCA) and Linear Discriminant Analysis (LDA) for data reduction and Multilayer Perceptron (MLP) to predict student’s performance on Learning Environments Systems (LMS). LMS are web-based educational software that provides classes and online assignments, and generate a large volume of data on student’s interactions in the environment. The method proposed, with the data used in this study, obtained better results in the prediction of the academic performance of students when compared to other methods approached in the literature.