Research of methods and algorithms for protecting network resources from spam

Abstract

This review investigates state-of-the-art methods and algorithms for protecting network resources from spam, with a focus on e-mail and network protocol spam in corporate networks. Approaches to filtering unwanted messages are reviewed, including rule-based filtering (e.g., keywords, heuristic rules), machine learning techniques (naive Bayes, decision trees, neural networks), blocking by IP addresses and accounts, and the use of blacklists (DNSBL, URI blocklists, etc.). The advantages and limitations of each approach are described, as well as scenarios of their application in a corporate environment. A comparative table of methods is given, indicating their applicability in corporate mail systems. The article is structured as an academic review: it contains an introduction, thematic sections on filtering methods, machine learning algorithms, blocking technologies and blacklists, a comparative analysis, a conclusion and a list of references. The latest 2020-2024 research publications, reports and standards (IETF, NIST) are used to support the conclusions.