Exploring the role of machine learning in Email Filtering
Haifa Khaled Alanazi, Rahaf Hamoud R. Al-Ruwaili · 2022 International Conference on Business Analytics for Technology and Security (ICBATS) · 2022
Spam, commonly known by various names as spam, commercial, or junk email, is the process or practice of sending some unsolicited bulk email messages to a randomly selected group of recipients, often to promote the sale of something. Phishing messages try to get your private and personal information by disguising it as legitimate messages from multiple trusted sources. They take you to many fraudulent websites, to fill in and use the private information that can be used for identity theft and many other things. Spam spreads across the Internet quickly because the value and transaction cost of electronic communications is much lower than that of any other and different type of communication. Spam is more than just annoying. It can be dangerous, especially if it is part of a phishing message. There are many agents used to filter spam through different methods to identify and mark an incoming message as spam, ranging from whitelisting, mail header analysis, content scanning, etc. With all the spam every day. This is not because the filters are not powerful enough, but rather because of the total and rapid adoption of all new technologies by spammers and the inflexibility of spam with changes. Therefore, modern machine learning algorithms offer many huge possibilities to help many organizations extract business value from the wealth and amount of data available today. All developers when starting the machine learning process rely on their knowledge and experience with statistics and the creation of many successful models. In this paper, we look at machine learning spam filtering strategies currently in use, as well as a few in development. We have discussed the use of three commonly used machine learning algorithms, namely C 4.5 decision tree, Naïve Bayes, and Multilayer Perceptron.