Extracting and Aggregating False Information from Microblogs
Naoaki Okazaki, Keita Nabeshima, Kento Watanabe, Junta Mizuno, Kentaro Inui · 2013
During the 2011 East Japan Earthquake and Tsunami Disaster, we had found a number of false information spread on Twitter, e.g., “The Cosmo Oil explosion causes toxic rain. ” This paper extracts pieces of false information exhaustively from all the tweets within one week after the earthquake. Designing a set of linguis-tic patterns that correct false information, this paper proposes a method for detecting false information. More specifically, the method extracts text passages that match to the correction patterns, clusters the pas-sages into topics of false information, and selects, for each topic, a passage explain-ing the false information the most suitably. In the experiment, we report the perfor-mance of the proposed method on the data set extracted manually from Web sites that are specialized in collecting false informa-tion. 1