An Analytical Study of Rumoured Tweets by Using Twitter Data

Abstract

Earlier when the internet was not there, rumours were spread by word of mouth technique but in this era of technology where we have social networking sites like twitter, rumours can be spread easily and quickly and a situation of panic can arise. Twitter is an American online news and social networking service on which users finds the latest news and world events faster. It is used for communication, interaction withpeople, announcement of event etc. from breaking news to sports, politics and everyday interests, one can find this service very addictive and an easy way to gather information about a certain event. Businesses can also use it to build their own brands and for marketing. But the founders of twitter like jack Dorsey forgot one thing that every coin has two sides. While twitter is a great way to interact with the masses, it is also a home of spammers. Spamming is a very common thing on twitter. Spammers create twitter accounts to perform a variety of tasks like posting links with unrelated tweets and the speed at which these fake and malicious misinformation spread on twitter in a real-time emergencies always causing a huge flood of tweets on twitter. In this paper, we demonstrated an analytical study of those rumoured tweets by twitter data. Using some of the rumoured tweets posted during the Chennai flood in 2015 and some non-rumoured tweets, we trained a classifier. The ability to track rumours and predict their outcomes have many applications for journalists, emergency services, and thereforehelp in minimizing the impact of false and fake information on this twitter platform.