Early Detection of Untrue Information in Social Networks

Abstract

Social networks have made spreading of information easier as information can now newlinebe shared by users without restrictions. Tools for quick and reliable veriand#64257;cation newlineof trustworthiness of posts on social networks are not readily available. In case of newlineinformation being untrue and sensitive, this fast and vast spread of information may newlinecause serious consequences. With social network becoming dominant channel for newlineinformation spread, studies on development of tools and techniques for quick and newlinereliable qualiand#64257;cation of trustworthiness of information has attracted researchers. newlineAttempts have been made for identiand#64257;cation of untrue information. Approaches newlineemphasizing on individual inand#64258;uence, community detection, epidemiological and newlinemachine learning have been proposed by researchers. Methods on identiand#64257;cation of newlineindividual inand#64258;uence are based on measuring or predicting user s ability to dissemi- newlinenate information, user s behaviour, interaction with other users and information newlinetransmission etc. Such methods also use number and type of interactions among newlineusers. Few of these methods classify users in multiple classes. The community de- newlinetection methods identiand#64257;es communities within social network based on link-based newlinecontents, and other attributes. Most researchers use a combination of attributes for newlineclustering users into communities. newline