Participation in Asynchronous Online Forums for Prediction of Learning Performance

The activity of students in online forums within an online learning platform is readily captured as a directed network graph whose structural and dynamic properties can be used to anticipate with good accuracy their level of performance in the future. In this chapter, we present and analyze a case study for this claim. Our data source is a complete record of the activity of students in forums, collected over two successive academic years of a computer networks course at the undergraduate level. We analyze these data with several social network analysis and machine learning algorithms, and identify as a result the best combination of features extracted from the graph for predicting the success or failure for individual students. The techniques are not computationally expensive, and can be used over a reduced dataset for having early estimates.