CSCL Chats Analysis and Scientific Threads Discovery

Chat is considered one of the most important and common tools that is widely used in cooperative education. Although different applications are used in Computer-Supported Collaborative Learning (CSCL) tools, only few applications help teachers analyze conversations automatically in order to obtain information capable of evaluating educational dialogues. This paper presents an effective mechanism that is capable of analyzing CSCL chats and extracting scientific threads automatically. The aim is to accomplish several tasks, including identifying the most important discussed threads. Based on Bakhtin's ideas and Trausan-Matu's polyphonic model, used for the analysis of the chat content which resulted in the extraction of the most repeated words. In addition to providing teachers with results of analysis in the form of statistical tables, they can conclude with the most important scientific threads discussed and the number of their repetition in the conversation. Accordingly, the paper relies on the most common perception of scientific subjects and their storage in a database, through which words are categorized if they are in the scientific or otherwise field. We will notice that the application analyzes chat conducted for a group of students from the NOOR International School for discussion of scientific threads.