Leveraging dynamic fuzzification to mitigate fraudulent profiles on social networking platforms

Le résumé fourni par la source

Due to the huge popularity of Online social networks (OSNs), people tend to utilize these platforms for information sharing, broadcast messages, and tend to make social connections with other users. The immense popularity and open interface of OSNs make them vulnerable to serious security threats, for instance, Sybil profiles, which are created to propagate false information, victimize individuals, hate speech, online fraudulent activities, and so on. These Sybil profiles are destructive for its users as well as for the service providers, and therefore, identification and filtration of such profiles is essential. In the context of Sybils’ detection, various behavioral, relational, and statistical approaches are presented. However, these approaches suffer from some key limitations, such as evasion tactics from intelligent adversaries, feature manipulation, data availability, unrealistic assumptions, and high false-positive and false-negative rates. To address these aforementioned challenges, this study offers a novel fuzzy ontology to detect sybils using socio-technical concepts and SWRL (Semantic Web Rule Language). Existing ontologies are not sufficient to be directly applicable or reusable for the fake profile problem. Therefore, building a novel ontology allows us to modify the representation, particularly that follows standard semantic web rules, which eventually enhances the reusability and interoperability with other systems in the future. To achieve this, Fuzzy logic and rules-based reasoner SWRL are employed using pre-defined rules to infer whether the OSN profile is Sybil or benign. Experimental results attain a higher detection accuracy of the proposed approach by generating results near human perception.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.