Recognizing and Stopping Rumors Patterns in Social Networks

Abstract

Objectives: In this study, a proposed Colored Petri Net Model (CPNM) is used for recognizing and stopping rumors in Social Networks (SN). Methods/Analysis: Detecting and blocking rumors represent an open security issue in social networks. In response to this issue, the proposed CPNM is experimentally simulated on dataset consists of 863-newsworthy tweets collected from the trending topic #CharlieHebdo in Twitter. The performance of CPNM is analyzed and evaluated using Precision, Recall, and Accuracy metrics. In addition, the CPNM is verified against the Reachability as a major behavior property in Petri Nets. Findings: The practical results disclosed a superiority of the proposed CPNM in detecting accurately rumors patterns compared with other approaches in the literature. In addition, verifying the Reachability using Reachability Graph proved that detecting and blocking rumors tweets are reachable states according to the firing life-cycle of tokens. Novelty/Improvement: Detecting rumors in social networks in more accuracy and low False Positive Rate (FPR) as well as blocking its propagation over the Social Network.