Spam Detection Using Machine Learning Algorithms and AdaBoost Technique

- E-mail has become an integral part of every individual's life. All sorts of communication formal and informal are carried out through this channel. As it is relatively cheaper and reliable, it has recently become a target for marketing agencies and hackers. A large number of unsolicited emails, called spam, containing either malicious content or unwanted advertisements, are sent to the individuals on a daily basis. This makes it difficult for e-mail service providers to segregate spam from non-spam (ham) e-mails. Consequently, important e-mails are often lost in spam, causing losses consumers and wasting valuable resources. The trust of consumers is compromised when their private information is accessed. This is where the need for improved performance of machine learning classifiers is felt. The aim of this paper is to classify spam emails by using supervised machine learning classifiers with and without Adaptive Boosting, and to conduct a deep analysis by using performance metrics. A significant improvement in the accuracy of SVM and Decision Tree has been observed.