Rumor Detection on Twitter Using a Supervised Machine Learning Framework

Introduction

When it comes to people and connectivity online social networking sites like Twitter seem to have set their web wider than any other. Studies reveal that the swift information propagation potential and extensive reach to the masses has made twitter the top source of breaking news in most parts of the world (Rosenstiel, Sonderman, Loker, Ivancin & Kjarval, 2015). Amidst the news statements, misinformation or rumor also finds its route into the society and its widespread is one of the biggest challenges facing social media (The guardian, 2011). Rumors are ubiquitous and with vast public involvement, they have the capability to impose real damage to individuals, organizations, and the government. Viral rumors about individuals that condemn them for their actions may lead to hate campaigns and eventually harm their reputation. This may affect individual’s self-esteem and confidence level. Rumors accelerate the dynamic nature of share markets and consequently elaborate their effect on organizations (Time,2013). Sometimes misinformation about the outburst of a disease (Time,2014) might affect the tourism of a country and likewise other government sectors. Analysis of rumors led to its aspect of public participation through various perspectives, for example, political belief (Shin, Jian, Driscoll & Bar,2016), influence on markets (Cruz & Gomes, 2013; Yiwen, Guizhong & Zongping, 2000) and crisis management (Onook, Manish & Raghav, 2013). For analysis of misinformation, its detection is the preliminary step.