A Machine Learning-based Approach for Twitter Spammers Detection

The ever-increasing popularity of Social Networks offers unprecedented opportunities to aggregate people and exchange information, but, at the same time, opens new modalities for cyber-crime perpetrations. The spamming phenomenon, so spread-out in emails, is now affecting microblogs, and exploits specific mechanisms of the messaging process. The paper proposes an inductive-learning method for the detection of Twitter-spammers, and applies a Random-Forest approach to a limited set of features that are extracted from traffic. Experimental results show that the proposed method outperforms existing approaches to this problem