3-TIER HYBRID APPROACH FOR SMS FILTERING

Abstract

Short Message Service is one of the most using services in mobile phones. In daily lift, several spam messages, which could disturb mobile phone users, are received frequently. Unwanted messages could be delivered for advertisement, annoucement of promotional events and/or only for disturbing people. In this study, 3-tier hybrid message filtering architecture has been introduced to protect the mobile users from spam messages. In the first two steps, incoming messages are classified based on the white /black lists and are identified as legitimate SMS and spam SMS. If the number of incoming message is not in those lists, the message is examined according to their meaning and morphological features. The success rate of k-Nearest Neighbor and Naive Bayes algorithms is about 96%. The proposed architecture has been implemented on Android platform.