A Review on Spam Classification Using Both Content-Full and Content-Less Features

Short message service (SMS) security is a very important issue in today's time. In spite of the fact that instant messengers have become very popular, SMS is still used. SMS is prone to several attacks and is not yet marked as a safe means of communication. Security in SMS includes message authentication, user authentication, encryption and decryption of the message and spam detection. This paper provides a comparison of multiple techniques available for detecting spam in messages. It deals with the survey of all the security techniques available, to find the best options available today. The most commonly used methods for spam classification are using support vector machines (SVM), relevance feature discovery, k-nearest neighbours, artificial immune system, Naïve Bayesian classifier or by content-less features. SVMs achieve substantial improvements over the currently best performing methods and behave robustly over a variety of different learning tasks. Furthermore, they are fully automatic, eliminating the need for manual parameter tuning.