Efficient genetic algorithm for spam mail detection and classification

Abstract

E-mail is the quickest tool to convey information from one to others. Individual users and organizations have become more reliant on e-mails as technology advances. At the moment, all email inboxes are swamped with spam and these cause significant and diverse difficulties, such as the loss of crucial information and the theft of the recipient’s identity. Organizations may suffer significant losses. As a result, users cannot avoid spam emails, which come in various formats such as advertisements and messages. Spam filtering removes spam messages and prevents them from being accessed. This paper focuses on categorization of e-mails by using genetic algorithm. The proposed approach uses entropy to evaluate information gain, which is subsequently employed by the genetic algorithm classifier to choose important features from the spam database. The model performance is then assessed by using the testing dataset with all 57 features, a crossover probability of 0.3, and a mutation probability of 1.0 and obtained good accuracy.