SMS Spam Detection Using Simple Message Content Features
Ghulam Mujtaba, Majid Yasin · 2014
Short message services (SMS) spam is increasing as more people exchange SMS messages very frequently. It is desirable to be eliminated for a number of reasons. This work describes a mobile station based approach where the spam sms would be identified and removed as soon as it is received at the mobile device. Four features are derived from each sms message and using these features a trained machine learning algorithm can classify an unknown message to be spam or ham. These features are the size of the message and existence of frequently occurring monograms in the message, existence of frequently occurring diagrams in the message and message class. The performance of Naive bayes algorithm is shown to be better than other algorithms explored. The other algorithms