Sockpuppet Detection in Wikipedia Using Machine Learning and Voting Classifiers

Rafeef Baamer, Mihai Boicu

2025

Abstract

Sockpuppet accounts, deceptive identities created by individuals on social networks, present significant challenges to online integrity and security. In this study, we analyse various approaches for detecting Sockpuppet accounts through the computation of several distinct features and the application of seven different classifiers. To enhance detection accuracy, we employ simple majority voting to aggregate the predictions from multiple classifiers, involving all seven classifiers, or only best five or three of them. This approach allows us to leverage the strengths of different classifiers while mitigating their individual weaknesses. Our experimental results show a significant improvement over previous studies, achieving an accuracy rate of 88% with 87% precision. Additionally, our experiments highlight the critical importance of feature engineering, demonstrating how carefully selected features directly influence classification performance. The findings also emphasize the value of human-in-the-loop involvement, where iterative feedback refines the models and improves their predictive capabilities. These insights offer meaningful contributions toward strengthening the security and integrity of online social networks by enabling more accurate and robust detection of Sockpuppet accounts.

Paper Citation

in Harvard Style

Baamer R. and Boicu M. (2025). Sockpuppet Detection in Wikipedia Using Machine Learning and Voting Classifiers. In Proceedings of the 10th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS; ISBN 978-989-758-750-4, SciTePress, pages 39-47. DOI: 10.5220/0013210100003944

in Bibtex Style

@conference{iotbds25,
author={Rafeef Baamer and Mihai Boicu},
title={Sockpuppet Detection in Wikipedia Using Machine Learning and Voting Classifiers},
booktitle={Proceedings of the 10th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS},
year={2025},
pages={39-47},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013210100003944},
isbn={978-989-758-750-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 10th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS
TI - Sockpuppet Detection in Wikipedia Using Machine Learning and Voting Classifiers
SN - 978-989-758-750-4
AU - Baamer R.
AU - Boicu M.
PY - 2025
SP - 39
EP - 47
DO - 10.5220/0013210100003944
PB - SciTePress