Twitter 是一個快速成長的社群媒體之一，在 Twitter 中的每一則訊息被稱為「tweet」，限於140個字元，因為眾多的使用者來分享他們的觀察，tweets 常常成為了報導真實世界中事件的即時訊息，特別是災害發生時，使用者常利用 Twitter 來報導何處需要幫助，這樣的 tweets 成為另一個潛在的資訊來源，並可支援災害救援。為了使用這些災害訊息容易被查詢、利用、和散佈，本文報告由 tweet 擷取災害救援資訊，並定位於線上地圖的方法和經驗。本文是以日本311地震後一個月內的 tweets 為研究對象，收集 tweets 之後，即進行具名實體辨識 (Named Entity Recognition) 將 tweets 中所提及的地名辨識出來，接著利用語彙的共現 (co-occurrence) 建立災害救援關鍵字的觀念階層(concept hierarchy)，基於觀念階層可以整理出三種三元組結構以擷取災害救援資訊，災害救援資訊擷取後存入空間資料庫中，可便於存取且顯示於地圖，如 Google Map，後端使用者 (end users) 因此可以在地圖上瀏覽災害救援資訊。

Twitter is a rapidly growing social media in recent years. The brief text sent viaTwitter is called as “tweets”, which is less than 140 characters. Since massive peoplepost tweets to report their observations, tweets are often real-time sources reflecting avarious kind of real-world events. In particular, people often post what/where they needwhen a disaster happens. These disaster tweets are potential information to supportdisaster relief. This study reports our experiences in extracting disaster informationfrom tweets via natural language process. While collecting tweets, we use NER (NamedEntity Recognition) technique to extract names of place and organization, and thenapply co-occurrence analysis to develop concept hierarchies of disaster keywords.Based on these concept hierarchies, we build three types of triple-like structures forextract disaster relief information. Since the disaster relief information is stored inspatial database, it can be retrieved easily, and displayed on the map such as GoogleMap. End users, thus, can browse the disaster relief information on the map.