Automatic Structure Generation of CSCL Discussion based on NLP Semantic Analysis

T. Tsurugi, M. Maejima, and Y. Tamura (Japan)


Computer supported collaborative learning, Content analysis, Natural language processing, Artificial intelligence


This paper proposes a method to automatically generate a structure of statements in CSCL activity. As a pre-process, NLP (natural language processing) is applied in order to clarify peer-to-peer inter-statement relationship of “Affirmative”, “Negative” and “Question”. Given this information, the authors developed a function to generate a whole structure of discussion. Since the overall structure cannot be predefined, Prolog language is utilized to identify a structure of the given discussion. With use of the proposing method, another learner group is able to understand and refer a discussion log of preceding learner groups.

Important Links:

Go Back