Performance Metrices of Different Machine Learning Algorithms

Abstract

E-mail is one of the most significant components or applications of the internet. Without e-mail, the internet would be useless. E-mail is a convenient and speedy method of communication. Email is a common way for people to exchange personal and professional information. Spam emails and non-spam emails are the two forms of emails. The purpose of spam email detection is to distinguish spam from valid e-mail. Spam is typically defined as uninvited or unwelcome email messages. The majority of spam emails contain viruses or other dangerous software that can cause computer and network malfunctions, use network bandwidth and storage space, and cause email servers to slow down. Because of the increase in the abundance of information, spam has become a big issue for internet users in recent years. Email phishing is detected and prevented using text mining techniques. Phishers trick people into providing personal information by sending fake emails and visiting fraudulent websites. Because of the increase in the volume of unwanted emails, there is an urgent need for the development of more dependable and robust email systems. Every day, email users receive hundreds of spam emails with unique content from anonymous addresses generated by robot software agents. This paperwork focusses to compare the accuracy, precision, recall, and F1 of the following algorithms: KNN, SVM. Nave bias and Random forest using emails as input.