Extracting and classifying typhoon disaster information based on volunteered geographic information from Chinese Sina microblog

摘要

The notion and application of volunteered geographic information occur and rise rapidly in recentyears with the thriving of social media in China such as SinaMicroblog, which is one of the mostactive social network sites. Many researches on natural disasters like flood, earthquake, and forestfires leverage social media like Twitter, Flickr, or YouTube, but few studies focus on typhoondisaster based on Sina Microblog even though typhoon disaster batters southeast coastline ofChina every year. This study proposed a method to extract and classify typhoon disaster informationfrom Sina Microblog. KNN (k-nearest neighbors) algorithm is implemented for microblogclassification in order to extract useful information about the real hazards, and an experiment isconducted to tune the parameters in KNNby comparison of outcomes of social media data analysisand the real typhoon situation. The result shows thatmore than 70% microblogs are classifiedcorrectly.After the classification,wecarried out spatial temporal analysis tomapthe disaster situation.It shows that the spatial distribution ofmicroblog messagemeancenters about typhoon hasregular variation along with the typhoon path. It can be confidently concluded that SinaMicrobloghas some potential prospects for estimating the typhoon disaster situation.