Spam Classification Using Genetic Algorithm


E-mail is the fastest way to exchange messages from one place to another across the world, the increased use of e-mail led to increase received messages in the mailbox, where the recipient receives many messages including those that cause significant and different problems such as stealing identity of recipient, losing of essential information causing losses to companies in addition to the damage to the network. These messages are so dangerous that the user is unable to avoid them especially as they take different forms such as advertisements and others. These messages are known as unwanted messages. In order to remove these spam messages and prevent them from being accessed, filtering is used. This paper aims to enhance the e-mail spam filtering by suggesting genetic algorithm classifier as a single objective evaluation algorithm problem to generate the best model to be used for classifying the e-mail messages in high accuracy. The first step in the proposal is applying normalization. The second is feature selection which is implemented to choose the best features, the third step is using genetic algorithm classifier as single objective evaluation algorithm that deal with one objective. The experimental results showed that the proposed system provides good accuracy in the first experiment (88%) and better accuracy in the second experiment (94%) and third experiment (95%).