Volume 18, No. 4, 2021
Email Spam Detection & Prediction Using Deep Clustering And Multi-Regression Models
I V S Venugopal , D Lalitha Bhaskari , M N Seetaramanath
Abstract
The growth in the digital form of information exchange have also cultivated the demand for higher security of the information and the information have become more sensitive for exchanges. The primary and most preferred form of the digital communication is email and majority of the attackers attempt to sabotage this communication method in the form of spam emails. Recognition of email spam is been the focal point of examination for more than 10 years at this point and numerous autonomous analysts and associations are attempting to assemble the most powerful type of spam filters in email servers. Throughout the investigation, this work revealed that the current spam separating, or recognition techniques are exceptionally time complex and a greater part of the cases are over fitted, hence conventional intentions is profoundly unimaginable. Additionally, the spam email location techniques are far away from the prescient recognition of the emails into ham or spam classes. Consequently, this work proposes two novel techniques for email location and expectation. The first deep clustering method demonstrates the email categorization process with 99.6% accuracy and the second method demonstrates nearly 99% accuracy for the prediction of the email spams. The proposed methods together are one of the bench marked research outcomes in this domain of the research for making the email based communications a more desirable method.
Pages: 70-85
Keywords: email spam, ham, email categorization, deep clustering, multi-regression