Research on Spam Filters using: SVM, Naïve Bayes, and KNN
Yiting Wang · Advances in computer science research · 2023
Email becomes a main way for people to communicate or send information to each other.However, spammers send people unwanted and harmful information using emails.Therefore, useful email filtering needs to be used for our email.This paper shows a comprehensive review and comparative concept of various spam filtering techniques by highlighting their strengths, weaknesses, and performance.The study focuses on three prominent approaches: K-Nearest Neighbors (KNN), Naïve Bayes, and Support Vector Machines (SVM).A large dataset of emails is used to determine how well each classifier performs.The testing set and the training set are two separate portions of the dataset.The computation of a number of performance metrics will be used.The performance metrics includes the precision, accuracy, f1-score, and recall of the specific filter.The analysis's findings show each technique's advantages and disadvantages.SVM exhibits great precision and accuracy but may be susceptible to parameter tuning and feature selection.KNN achieves competitive results with a straightforward implementation but can suffer from scalability issues.Naïve Bayes, despite its simplistic assumptions, performs well too.