Spam detection in SMS based on feature selection techniques

Abstract

Short Message Service (SMS) has become the most effective and efficient means of communication. The popularity for SMS is due to their ease of use. But with the advent of this media, the problem of SMS spamming has increased. These are undesired and illegal messages which cause a lot of inconvenience to the users. So in order to get rid of these spam messages, classification algorithm along with feature selection techniques is incorporated. This technique helps us to choose better features and provides a better accuracy. The irrelevant and redundant attributes that do not contribute to the accuracy of the model are removed using the feature selection techniques. The complexity of the model is reduced as only fewer attributes are desired and the simpler model is easy to understand. A comparative study of different algorithms has also been studied in this work.