Extracting Non-Situational Information from Twitter During Disaster Events

Extracting Non-Situational Information from Twitter During Disaster Events

Poonam Sarda (Department of Computer Science and Engineering, Government Engineering College Bikaner, Bikaner, India) and Ranu Lal Chouhan (Department of Computer Science and Engineering, Government Engineering College Bikaner, Bikaner, India)
Copyright: © 2017 |Pages: 9
DOI: 10.4018/jcit.2017010102

Abstract

Micro blogging sites have become important forums for discussion during disaster events in which Twitter has become one of the important source of real time information. Millions of tweets are posted during disasters, which include not only information about the present situation or relief efforts, but also the emotions or opinion of the masses. Much research has been conducted on extracting situational information from tweets during disaster. However, according to current knowledge, there has not been any prior attempt to study the non-situational tweets posted during disasters, such as those which express the emotions/opinions of the people, political and governmental views, raising charities and event analysis. In this study, the authors characterized the non-situational tweets posted during recent disaster events, the Nepal Earthquake and the Gurudaspur Terrorist attack. They developed a classifier to categorize various types of non-situational tweets into a set of fine-grained classes utilizing state-of-the-art machine learning technique. This system also helps in filtering out communal tweets which can make worst the situation by disrupting communal harmony during certain disaster events.
Article Preview

1. Introduction

Twitter, a micro-blogging social networking website has a large and rapidly growing user base. Rich bank of data is provided by the Twitter in form of 'tweets' which must be written within 140 characters. The experiment in Go et al. (2009) found that the average length of tweets is 14 words or 78 characters. Some of the applications which rely on Twitter data are analysis of disasters (Sen et al., 2015; Brynielsson et al., 2013), detection of diseases (Dai and Bikdash, 2015; Grover and Aujla, 2015; Grover et al., 2014; Aramaki et al., 2011), political elections (Asur and Huberman, 2010), movie review (Bollen et al., 2011) and stock market (Tumasjan et al., 2010). In tweets, abbreviations, orthographic mistakes, emoticons and hash tags are frequently used to express the message in few words. There have been a lot of researches on analyzing tweets posted during disasters, and most of the prior studies have focused on extracting situational information, that is, information which helps to gain a high-level understanding of the circumstance (Sarter and Woods, 1991; Vieweg et al., 2010). For instance, several studies have been done to develop classifiers for differentiating situational tweets from other non-situational tweets (Sen et al., 2015; Imran et al., 2013), while some studies not only attempted to summarize situational tweets (Sen et al., 2015; Nguyen et al., 2015) in English but also in different languages like Hindi (https://en.wikipedia.org/wiki/2015_Gurdaspur_attack; Sharma et al., 2015).

Complete Article List

Search this Journal:
Reset
Open Access Articles
Volume 21: 4 Issues (2019): Forthcoming, Available for Pre-Order
Volume 20: 4 Issues (2018): 2 Released, 2 Forthcoming
Volume 19: 4 Issues (2017)
Volume 18: 4 Issues (2016)
Volume 17: 4 Issues (2015)
Volume 16: 4 Issues (2014)
Volume 15: 4 Issues (2013)
Volume 14: 4 Issues (2012)
Volume 13: 4 Issues (2011)
Volume 12: 4 Issues (2010)
Volume 11: 4 Issues (2009)
Volume 10: 4 Issues (2008)
Volume 9: 4 Issues (2007)
Volume 8: 4 Issues (2006)
Volume 7: 4 Issues (2005)
Volume 6: 1 Issue (2004)
Volume 5: 1 Issue (2003)
Volume 4: 1 Issue (2002)
Volume 3: 1 Issue (2001)
Volume 2: 1 Issue (2000)
Volume 1: 1 Issue (1999)
View Complete Journal Contents Listing