A Novel Hybrid Algorithm to Classify Spam Profiles in Twitter

Spam profile detection problem is a huge threat to social networks and poses severe challenges to safeguard the identity of users. The consequences due to the existence of spam profiles are alarming. Though several techniques have been proposed in history to identify the spam profiles, the usability of these methods is very limited due to the evolving nature of spammers. A method devised for a spammer strategy may not work when the spammers change their identity and behavior. Hence, there is a need to develop spam detector systems that are robust and work effectively even when the spammers‟ strategies are evolving. In this paper, a unique hybrid wrapper based technique to detect spam profiles in the online social network, MWOA-SPD is proposed. The Whale Optimization Algorithm (WOA) is integrated with the Salp Swarm Algorithm (SSA) to achieve better classification accuracy with a minimal subset of features. The exploration technique of WOA is replaced with the position updating mechanism of SSA to diversify the search and to get rid of the limitation of WOA. A dataset was extracted from Twitter and used as a benchmark to evaluate the performance of the proposed method. The findings of the results show that the proposed method yields competitive results compared to the existing ones.