TRENDREADER (TR)を用いた災害・危機に関する言語資料体の解析にもとづく災害・危機事象の将来展開予測手法の開発 : 新潟県を襲った近年の2つの地震災害を例にして

概要

論文の詳細を見る
Emergency Operation Center should have intelligence for prediction not only about probability and scale of natural hazard, but also tendency of social phenomena of disaster for proactive disaster response. However, unlike natural hazard estimation, there is no method or system that could predict social trend. In this paper, we propose a prediction method of social phenomena tendency based on analysis of disaster corpus using TRENDREADER (TR).The TR system automatically extracts meaningful keyword sets from digital text datasets related to disasters, and visualizes a "big picture" of whole situation. Our method consists of the following two steps; 1) to identify keywords among documents about possibly significant disaster social phenomena and 2) to apply trend of the keywords in stock chart model and to provide daily summary as well as weekly summary.
地域安全学会の論文