Probability-based Naïve Bayes Algorithm for Email Spam Classification

Abstract

Email spam, which is a type of e-spam, is one of the most common internet problems. Email is the standard mode of communication for sharing vital and official information. Most institutions and businesses prefer email to all other forms of communication because it is the most cost-effective, simple to use, easily accessible, official, and dependable. It's commonly utilized since it ensures that the information submitted is kept private. The majority of spam emails are sent for commercial objectives, and others may contain virus links that direct users to fraudulent websites. However, there are drawbacks to this reliable and simple mode of communication, as many people abuse it by sending unwanted and pointless messages for their benefit. These unwanted emails generate problems for the average user, such as filling the inbox with undesirable emails, making it difficult to find valuable emails, and even causing one to miss over vital and beneficial communications as a result of all these unwanted emails. As a result, a powerful email spam detector is required, one that can filter a large number of spam emails with increased accuracy while ensuring that real emails are not screened as spam. Ham consists of emails that are legally legitimate messages that people can accept. Unwanted emails that a user wants to delete are referred to as spam. The goal of this study is to use an improved and efficient classification algorithm to classify spam and ham emails. The goal of this study is to enhance the accuracy of classifying emails into two groups with minimal training. This study uses the Nave Bayes (NB) classifier to ensure that the requirements are met with minimal training and that the findings are more accurate than previous methods.