Malicious Accounts Detection Based on Short URLs in Twitter

Abstract

The popularity of Social Networks during the last several years has attracted attention of cybercriminals for the spreading of spam and malicious contents. In order to send spam messages to lured users, spammers creating fake profiles, leading to fraud or malware campaigns. Sometimes to send malicious messages, cybercriminals use stolen accounts of legitimate users. Nowadays they are creating short URLs by the short URL service provider and posted on to friends board. Lured users unknowingly clicking on these links, then they are redirected to malicious websites. To control such type of activities over Twitter we have calculated a trust score for each user. Based on the trust score, one can decide whether a user is trustable or not. With usage of trust score we have got an accuracy of 92.6% and F-measure is 81% with our proposed approach