Presentation | 2014-06-14 Early Detection of Disasters with Contextual Information on Twitter Shota SAITO, Yohei IKAWA, Hideyuki SUZUKI, Akiko MURAKAMI, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Twitter, a recently growing micro-blogging service, offers opportunities to analyze real-time events by its real-time nature. Particularly, it is highly valuable if real-time disasters in a real world can be detected by Twitter. However, we have to search for the disaster on Twitter manually, or we cannot know the details or the contexts of the disaster if we use previous porposed event detection methods. Therefore we propose a method for an early detection of disasters with contextual information. In this study, we assume that huge disasters make a topic which is composed of several words, and that the expression of tweets mentioning that disaster diverge. Based on the assumputions, we propose to make a graph of cooccurence of the words appearing in Twitter, and to detect real disasters by independent source measure, which measures how much the expressions between each tweet mentioning a same topic diverge. We demonstrate our technique on real data from Twitter and show that our method can detect reasonable disasters before media reports. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Social Media / Event Detection / Graph / Indenpendent Source Measure |
Paper # | NLC2014-2 |
Date of Issue |
Conference Information | |
Committee | NLC |
---|---|
Conference Date | 2014/6/7(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Natural Language Understanding and Models of Communication (NLC) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Early Detection of Disasters with Contextual Information on Twitter |
Sub Title (in English) | |
Keyword(1) | Social Media |
Keyword(2) | Event Detection |
Keyword(3) | Graph |
Keyword(4) | Indenpendent Source Measure |
1st Author's Name | Shota SAITO |
1st Author's Affiliation | Graduate School of Information Science and Technology, University of Tokyo() |
2nd Author's Name | Yohei IKAWA |
2nd Author's Affiliation | IBM Research - Tokyo |
3rd Author's Name | Hideyuki SUZUKI |
3rd Author's Affiliation | Graduate School of Information Science and Technology, University of Tokyo |
4th Author's Name | Akiko MURAKAMI |
4th Author's Affiliation | IBM Research - Tokyo |
Date | 2014-06-14 |
Paper # | NLC2014-2 |
Volume (vol) | vol.114 |
Number (no) | 81 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |