Systematic Review of Machine Learning and Deep Learning Approaches for Spam Detection in Online Social Networks
Impact in Computics · 28.06.2026 · 2 · 1-15
DOI: 10.65500/computics-2026-004 ↗Kurz gefasst
Systematische Übersicht in der Zeitschrift Impact in Computics, online erschienen am 28.06.2026. Verfasst von Abdu Ibrahim und 4 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Malicious activities such as phishing, disinformation, and bot attacks for spamming in online social networks (OSNs) have become an increasing problem, and online spam detection is an ongoing challenge to keep users safe and OSNs healthy. However, classical machine learning (ML) algorithms like Random Forest (RF), Decision Trees (DT), k-Nearest Neighbors (KNNs), and Naïve Bayes (NB) have various limitations, including class imbalance, overfitting, and scalability, in the context of OSN environments. Emerging deep learning (DL) algorithms, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Long Short-Term Memory (LSTM) networks, are more tolerant of high dimensionality and temporal data. With recent developments, Graph Neural Networks (GNNs), attention mechanisms, and transformers, along with ensemble learning, are being used to model structural and semantic spam patterns. It presents the results of a systematic review of 25 peer-reviewed papers published between 2020 and 2025 and discusses them in terms of approaches, performance assessment, and limitations. The review found that hybrid DL approaches generally achieved better spam detection performance than traditional machine learning (ML) methods, although at the cost of increased computational complexity and resource requirements. However, since 2020, no systematic synthesis has examined the trade-offs between the performance and efficiency of hybrid models. We observed that hybrid models that integrate GNNs, attention mechanisms, and optimization techniques have great promise for future research and can be used to provide real-time and scalable spam detection for multiple social platforms.
Zitieren
APA
Ibrahim, A., Ya’u, B. I., Zambuk, F. U., Adamu, S., & Egere, A. N. (2026). Systematic Review of Machine Learning and Deep Learning Approaches for Spam Detection in Online Social Networks. Impact in Computics, 2, 1-15. https://doi.org/10.65500/computics-2026-004
Vancouver
Ibrahim A, Ya’u BI, Zambuk FU, Adamu S, Egere AN. Systematic Review of Machine Learning and Deep Learning Approaches for Spam Detection in Online Social Networks. Impact in Computics. 2026;2:1-15. doi:10.65500/computics-2026-004
BibTeX
@article{Ibrahim2026,
author = {Ibrahim, Abdu and Ya’u, Badamasi Imam and Zambuk, Fatima Umar and Adamu, Shehu and Egere, Augustine Ndudi},
title = {Systematic Review of Machine Learning and Deep Learning Approaches for Spam Detection in Online Social Networks},
journal = {Impact in Computics},
year = {2026},
volume = {2},
pages = {1--15},
doi = {10.65500/computics-2026-004}
}
EUVIMED