Short Message Service (SMS) has become a common communication mode, specifically in Pakistan. In 2009, Pakistan had the largest text messaging growth in Asia Pacific. The reduction in the cost of SMS services by telecom companies has enabled the increased use of SMS. The use of SMS in financial transactions e.g., bank transactions, mobile money, bill payments, notification of due dates, as well as one-time PINs (OTP) is extremely common. Due to the frequency and familiarity of these SMSes, attackers sometimes send similar messages as spam or fraud, seeking to either extract some private information or to defraud them by transferring money to some third-party scam, from legitimate traffic.

This project focuses on exploring the landscape of fraudulent and spam SMS messages in the SMS messages universe in Pakistan. For this purpose, SMS data has been collected from multiple sources over a period of five months. Further, this project builds on the data and develops a machine-learning algorithm to identify and tag fraudulent and spam messages in the data corpus.