Improving Static SMS Spam Detection by Using New Content-based Features

摘要

As the use of mobile phones grows, spams are becoming increasingly common in mobile communication such as SMS, calling for the research on SMS spam detection. Most of the existing detection techniques for SMS spams are have been mostly adapted from other contexts such as emails and the web but ignored some unique characteristics of SMS in spam detection. Moreover, spammers are constantly developing more sophisticated tactics, rendering previous features and methods for spam detecting no longer effective. In this paper, we propose to incorporate different new content based features to improve the performance of SMS spam detection. The effectiveness of the proposed features is empirically validated using multiple classification methods. The results demonstrate that the proposed features can improve the performance of SMS spam detection.