Detecting Fake Users on Social Media with a Graph Database
Abstract
Social media has become a major part of people’s daily lives as it provides users with the convenience to connect with people, interact with friends, share personal content with others, and gather information. However, it also creates opportunities for fake users. Fake users on social media may be perceived as popular and influential if not detected. They might spread false information or fake news by making it look real, manipulating real users into making certain decisions. In computer science, a social network can be treated as a graph, which is a data structure consisting of nodes being the social media users, and edges being the connections between users. Graph data can be stored in a graph database for efficient data analysis. In this paper, we propose using a graph database to achieve an increased scalability to accommodate larger graphs. Centrality measures as features were extracted for the random forest classifier to successfully detect fake users with high precision, recall, and accuracy. We have achieved promising results especially when compared with previous studies.
Copyright (c) 2021 Yichun Zhao

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Authors contributing to the Artbutus Review agree to release their articles under the Creative Commons Attribution-Noncommercial 4.0 Unported license. This licence allows anyone to share their work (copy, distribute, transmit) and to adapt it for non-commercial purposes provided that appropriate attribution is given, and that in the event of reuse or distribution, the terms of this license are made clear.
Authors retain copyright of their work and grant the journal right of first publication.
Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.