Groups Allocation Based on Sentiment-Epistemic Analysis in Online Learning Environment

Collaborative learning methods in an online learning environment, encourages students to interact actively among themselves in a work group. Instructors or lecturers need to combine potential students to work together as a group, but this task is not easy since the characteristics of a student are sometimes not explicitly known. In this preliminary research, we propose a solution to answer this problem. Our methodology is composed in three steps. It begins with the sentiment analysis process with a textual history of online conversation or discussion. The next step is to classify the text into one of predefined epistemic groups. Further, we visualize the model in an epistemic network graph which is based on singular value decomposition. The group allocation is built based on k-means clustering. The case study in this paper is related to information technology-based subjects, and thus we classify our sentiment-epistemic analysis in three collaborative aspects, i.e.: project management, attitude and technology affinity. Our results show that by combining sentiment-epistemic analysis and k-means clustering, a holistic group allocation can be produced which would be beneficial in a collaborative learning environment.