2026-08-01T02:23:06.856Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:9815 2024-04-04T11:01:16Z 3

Efficient spam email classification logistic regression model trained by modified social network search algorithm 2024 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/9815 https://www.sciencedirect.com/science/article/pii/B9780443132681000108 B. Radomirovic A. Petrovic M. Zivkovic A. Njegus N. Budimirovic N. Bacanin Due to the frequent interruptions it causes during work or personal time, spam is a real annoyance for email users. As a result of their effectiveness and often high classification accuracy, machine learning methods are frequently employed as the core of spam detection systems. On occasion, valid emails are designated a spam label; more commonly, though, a few spam emails land in the user’s inbox and appear to be legitimate. By using improved social network search metaheuristics to train a logistic regression model, this paper suggests a unique approach for email spam detection that addresses the inadequacies of the available methods. The created approach has been evaluated against a publicly available high-dimensional spam benchmark dataset (CSDMC2010), and thorough trials have demonstrated that the model handles high-degree data efficiently. The suggested model achieves higher classification accuracy, as demonstrated by a comparison with existing state-of-the-art spam detection methods. bookPart Elsevier 39 55 10.1016/B978-0-443-13268-1.00010-8 Chapter in Advanced Studies in Complex Systems, Computational Intelligence and Blockchain in Complex Systems