Twitter利用者の動的定量化手法の提案-Twitterの災害時活用に向けて-

書誌事項

タイトル別名
  • Proposal of Dynamic Quantification Analysis Method of Twitter Users : Toward the Utilization of Twitter at the Time of Disaster
  • Twitter リヨウシャ ノ ドウテキ テイリョウカ シュホウ ノ テイアン : Twitter ノ サイガイジ カツヨウ ニ ムケテ

この論文をさがす

抄録

application/pdf

The authors have previously addressed the analysis of Twitter to automatically extract valuable information sent by humans with high real-time property. Since a lot of information sent by automatioc posting programs called "bot" has mixed in Twitter, currently a method to remove bot's voice has been developed. From our past studies, the identification of bots with satisfactory accuracy was confirmed by the filtering method. Based on the previous efforts,this paper proposes entropy based dynamic quantification analysis method of Twitter users toward the utilization of twitter at the time of disaster.If a disaster occurs, long-term data on Twitter will not be available because valuable information for the disaster should be collected as soon as possible. Therefore in this paper, the usual method is improved by changing the range of data set for analysis dynamically. This paper performs some experiments with the real Twitter data around the Great East Japan Earthquake, and shows the effectiveness of the proposed method.

収録刊行物

詳細情報

問題の指摘

ページトップへ