Machine Learning Approaches for Fake User and Spammer Detection: A Comprehensive Review and Future Perspectives
DOI:
https://doi.org/10.31033/ijemr.13.3.4Keywords:
Machine Learning, Spammer Detection, Fake User Detection, Fraud Detection, Deep Learning, Unsupervised Learning, Semi-Supervised Learning, Transfer Learning, Active Learning, Federated Learning, Explainable AI, Reinforcement Learning, Online SecurityAbstract
The rise of digital platforms has given way to a surge in fraudulent activities, including the creation of fake user accounts and the prevalence of spammers. These malevolent actions present significant challenges to the security and integrity of these platforms, necessitating effective detection and prevention measures. This paper offers an extensive review of machine learning (ML) techniques currently employed for fake user and spammer detection. The paper explores a range of traditional ML algorithms such as decision trees, support vector machines, and logistic regression, as well as more complex deep learning models like convolutional neural networks (CNN) and recurrent neural networks (RNN). It also examines unsupervised and semi-supervised learning strategies that can be used when labeled data is scarce. Furthermore, we discuss the key challenges in detecting fake users and spammers, including the dynamic nature of spamming tactics, evolving deceptive strategies, data imbalance, and privacy issues. We propose potential solutions to these challenges like transfer learning, active learning, federated learning, and privacy-preserving ML techniques. The paper concludes with an exploration of emerging technologies such as explainable AI and reinforcement learning and their potential to enhance detection system performance and interpretability. It also provides insights into promising future research directions in this critical area.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2023 Farheen Siddiqui, Mohammad Suaib
This work is licensed under a Creative Commons Attribution 4.0 International License.