Sentiments in Social Context of Student Modelling

Abstract

Social learning analytics is an emerging discipline that offers new methods to explore data from online educational devices in order to obtain a better understanding of student behavior. As learning takes place in heterogeneous and complex online environments, the incorporation of contextual information about the student has attracted major interest. For some time, most methods of student modelling have considered interactions as a key dimension of the student's social context, but only recently, automatic extraction software agents begin to tackle interactions in a non-exclusively quantitative way. In this paper, we propose the discovery of sentiments in online social interactions as an additional property for the modeling of students in order to produce a contextualized diagnosis when performing learning analytics. We propose answers to the questions of "Which are the sentiments in students context modeling?", "Why are they important for social learning analytics?", "How can we visualize them?"