Machine Learning Approaches for Email And IoT Spam Detection: Analysis and Challenges

Electronic mail, or email, has proliferated in today's society in a number of sectors, including educational institutions and business. There are two primary categories of emails: lawful ham emails and unsolicited or harmful spam emails. Spam emails are becoming a bigger problem since they not only take up users' time and computer resources but also jeopardise the security of sensitive data. Strong techniques for identification and filtering are vital, especially for internet / the web of Things firms, since the percentage of spam emails keeps rising. Of all the strategies used to combat spam, email filtering is one of the most important. Several machine instruction and deep learning techniques, such as Naïve Bayes, neural network training, decision trees, and random forests, have been investigated for efficient spam filtering. In this study, a thorough assessment and systematic classification of various machine learning methods for spam filtering on email and internet of things systems is carried out. Metrics like precision, precision, precision, and recall are used to evaluate various methods and provide important information about how they compare. Finally, this synthesis provides a plagiarism-free examination of machine learning methods used to prevent spam in emails and Internet of Things platforms. The text is easier to read and retains the main ideas of the original discussion when the number references are removed. The talk ends by summarising the most important discoveries and outlining possible directions for further investigation in this important area.