Natural Language Processing to Analyze Online Discussions: Fourth-Grade Small-Group Online Discussions in E-Books

• In Event: Learning Language and Literacy Online

Abstract

Effective moderating of online discussions is vital for online discussions. Yet, moderating of online discussions is notoriously labor-intensive for instructors. We collected small group, “book club,” discussion data from Slovenian fourth graders reading web-based eBooks, qualitatively coded it and then passed the data on to six groups of computer science students who developed Natural Language Processing (NLP) algorithms to automatically classify the discussion data, for instance as book-relevant/discussion-relevant. One group predicted book-relevance at 85%, in Slovenian postings and 95% in the English translations. Another group used context information to predict the categories of postings, as well as the category of the following posting. Our results suggest that NLP could assist teachers moderating online discussions, providing real-time notification of discussion problems.

Authors

• Glenn Gordon Smith, University of South Florida

• Slavko Žitnik, University of Ljubljana