An Enhanced Mechanism for Deep Learning based Spam E-Mail Detection

Abstract

In the world of the internet, Spam email is becoming the biggest issue. Financial organisations are influenced by spam emails, which also exacerbate individual email users' experiences. At present, email communication is becoming more frequent, making it more challenging to maintain its security and integrity and detect spam emails. The development of effective spam detection models has been an important advance in existing research, but adaptability and classification performance remain challenging as spamming techniques evolve. This presents an Enhanced Mechanism for Deep Learning based Spam E-Mail Detection. From the Enron Corpus, raw email data was collected. The e-mail classification stage includes mapping between the training and test sets using Convolutional Neural Networks (CNNs). This analysis describes the Sand Cat Swarm Optimisation (SCSO) algorithm, which employs a CNN to enhance accuracy and minimise loss. The described model distinguishes between undesired (spam) and genuine (non-spam) emails. Accuracy, Precision and Recall are the parameters used in this paper to measure performance. Results confirmed that the described model gives very accurate results.