A Comprehensive Review of Spam Detection Techniques: From Traditional Methods to Advanced Computational Systems
Maninder Kaur, Gagandeep Kaur · International Journal of Research Publication and Reviews · 2025
This paper thoroughly reviews the various computational models used in spam detection, highlighting the progression from basic techniques to more sophisticated approaches.Emphasis is placed on transformer-based architectures and supervised learning methods such as Naive Bayes, Random Forest, and Support Vector Machines (SVM), as well as hybrid models that combine supervised and unsupervised learning.Transformer models show a notable accuracy, often above 94%, but their requirement for large datasets and significant computational resources presents challenges.Traditional algorithms are accurate but face scalability issues, while hybrid models provide a balance by combining multiple approaches to address spam.This review also looks into dynamic rule generation systems integrated within email servers for real-time filtering, especially in resource-limited environments like IoT.Despite their effectiveness, these models face challenges in terms of energy efficiency, data privacy regulations, and computational load.The literature review utilizes reputable sources like IEEE Xplore, ACM Digital Library, and SpringerLink to ensure credibility.It also identifies research gaps and suggests directions for future advancements, particularly in creating energy-efficient, scalable, and privacy-preserving systems.This review paper systematically examines various computational models techniques employed in email spam detection, highlighting the shift from traditional approaches to advanced models.The study focuses on transformer-based architectures, supervised learning algorithms such as Naive Bayes, Random Forest, and Support Vector Machines (SVM), as well as hybrid models that integrate both supervised and unsupervised learning techniques.In addition to this, this paper explores the use of dynamic rule generation systems embedded within email servers, providing real-time spam filtering in resource-constrained environments such as IoT devices.These systems offer a scalable and cost-effective solution, although challenges related to computational demands, energy efficiency, and privacy regulation compliance (e.g., GDPR) remain significant.Addressing these limitations is crucial for the development of effective spam detection systems in embedded applications.