Hybrid Particle Swarm Optimization and Pegasos Algorithm for Spam Email Detection
L.., El Bakrawy · Russian Agency for Digital Standardization · 2019
Email is one of the most popular communication tools for most internet usersnowadays. It has become fast and an effective method to share and exchange information allover the world. Despite the great advantages of emails, its usage is facing problem which is spamemails. Spam emails are the huge presence of bulk and unsolicited emails which are expensive forthe companies, consume a huge amount of mail servers, network bandwidth and waste of time.Isolating and detecting these emails is known as spam detection. Many spam detection methodshave been proposed but there is still need to detect the email spam effectively with high accuracy.In this paper, hybrid particle swarm optimization and Pegasos algorithm, which is called (PSOPegasos) is proposed for spam email detection. Particle swarm optimization is employed as asearch strategy to determine the optimal parameters for Pegasos algorithm in order to achievehigher performance. The proposed algorithm has been applied on spambase dataset downloadedfrom UCI Machine Learning Repository. Experimental results demonstrate that the proposedalgorithm outperforms the performance of all the earlier proposed algorithms, considering theaccuracy, recall, precision and Fmeasure on the same dataset.