Content Based Email Spam Filtering and Detection using Hybrid Supervised Learning Approach
Najam U. Saquib, Waqar Khalid, Sahibzadi Annum Shaheen, Muhammad Asim, Naveed Ahmad · 2024
In today’s world, email is used widely for communication purposes globally. Email spam are unwanted emails that are sent to many recipients receivers. It is usually used for commercial purposes More likely spam enail is also used for malicious attempts to gain access to receiver computer/device. This article presents a novel hybrid architecture for spam emails filtering using mails content-based approach. For filtering, Word2Vec (W2V) word embedding procedure is used for feature selection with bio-inspired “Particle Swarm Optimization (PSO)” and “Support Vector Machine (SVM)” combined with “Convolutional Neural Network (CNN)”. Moreover, our proposed hybrid architecture is simulated in open-source tool and the performance is evaluated using standard confusion matrix. Afterw ards, the results are compared with various previous algorithms/models. Experimental results show that our proposed hybrid scheme enhances detection accuracy of spam emails.