Evaluating the Network Performance of the Ensembled-Based Veracity Architecture for Fake News Detection in Infrastructureless Social Networks

Infrastructureless social networks (ISNs) are created by the interconnection of spectral-constrained mobile devices. One such type of ISN are the mobile ad hoc networks (MANETs). One of the issues that consumers of content on ISNs face is the inability to detect fake news in messages sent through the network. To address the fake news detection issue in ISNs, the ensemble-based veracity architecture, an ensemble-based computational social system for fake news detection in infrastructureless social networks, has been proposed. Ensemble-based Veracity detects fake news using both the publisher’s credibility and the content of the news. To understand the effect that ensemble-based Veracity has on network performance, this work investigates the network performance of the ensemble-based Veracity architecture. Ensemble-based Veracity is fully evaluated using a MANET-based experimental design and simulation environment. The network performance results of the experiments on ensemble-based Veracity are thoroughly analysed, and all the observations are noted. According to the experimental results, the throughputs were 2,445,528 bps, 2,391,905 bps and 2,236,778 bps for 20, 50 and 100 nodes, respectively. The experimental results show that ensemble-based Veracity negligibly affects the throughput, queuing time, queue length and number of packets passed to the upper layers of the network and the network performance.