Portal de Periódicos da CAPES

Training Logistic Regression Model by Hybridized Multi-verse Optimizer for Spam Email Classification

2023; Springer International Publishing; Linguagem: Inglês

10.1007/978-981-19-6634-7_35

2367-3370

Miodrag Živković, Aleksandar Petrović, Nebojša Bačanin, Marko Djuric, Ana Vesić, Ivana Strumberger, Marina Marjanović,

Data Stream Mining Techniques

Spam emails pose a significant threat to end users, annoying them and wasting their time. To counter this problem, numerous spam detection systems have been proposed recently, where the most of the solutions have grounds in the machine learning algorithms, due to their efficiency in classification tasks. Unfortunately, existing spam detection solutions typically face low detection rate and generally have troubles in dealing with high-dimensional data. To address this problem, this paper suggests a hybrid spam detection approach by combining the logistic regression classifying model with the hybridized multi-verse optimizer swarm intelligence metaheuristics. The proposed approach was validated on a public benchmark dataset (CSDMC2010) and compared to other cutting-edge techniques. The obtained results indicate that the suggested hybrid approach outperforms other spam detection solutions included in the comparative analysis, by achieving the highest classification accuracy.