This paper proposes a method for detecting new concepts in social media using co-burst pattern mining technique. The target social media are viewers ’comments that attached to web videos, and Twitter ’s tweets, that are related to the East Japan Great Earthquake that happened on Mar. 11 in 2011. Our proposed method rst crawls viewers ’comments attachedto web videos, and extracts words from them. It selects motive words candidates from extracted word, and then counts the occurrence numbers of tweets that include the motive words. To detect new concepts, it analyzes burst correlations based on occurrence numbers of motive words over time using co-occurring pattern mining. By our method, new burst correlations between motive words triggered by social media are recognized as new concepts. For example, we could detect the new concepts (e.g. ”escape, nuclear plant ”) on Twitter after the earthquake. In this paper, we provide experimental results and discuss the effectiveness of our proposed method.
