Data Mining of the Concept "End of the World" in Twitter Microblogs
This work addresses the problem of predicting societal reactions to events for social media analysts, but it is incremental as it applies existing data mining methods to a specific dataset.
The paper analyzed Twitter posts about the 'end of the world' predicted for December 21, 2012, finding that frequent sets and association rules in the data can serve as predictive markers, with support peaking before the event and confidence dynamics indicating potential societal signals.
This paper describes the analysis of quantitative characteristics of frequent sets and association rules in the posts of Twitter microblogs, related to the discussion of "end of the world", which was allegedly predicted on December 21, 2012 due to the Mayan calendar. Discovered frequent sets and association rules characterize semantic relations between the concepts of analyzed subjects.The support for some fequent sets reaches the global maximum before the expected event with some time delay. Such frequent sets may be considered as predictive markers that characterize the significance of expected events for blogosphere users. It was shown that time dynamics of confidence of some revealed association rules can also have predictive characteristics. Exceeding a certain threshold, it may be a signal for the corresponding reaction in the society during the time interval between the maximum and probable coming of an event.