A Content-Based Approach for Detecting Smishing in Mobile Environment

Abstract

Rapid development in Information Technology has led to increased usage of smartphones. Smartphone users are storing their sensitive information like their user credentials, credit card, and debit card information in the mobile device. Moreover, mobile devices are constantly connected to the World Wide Web through packet data connection or Wifi which makes these devices prone to phishing attacks. Smishing is a combination of Sms and Phishing in which attackers target the mobile user through a text message sent to their mobile device. These text messages contain a link which will redirect the user to malicious websites. Many methods are proposed by researchers in past years to mitigate the smishing attacks which included SMS feature-based analysis, blacklisting techniques, and heuristic methods. But still, we don't have a method which reduces false positive results. Hence, We have proposed a novel method which will categorize the text message based on the SMS contents and URL behavior. SMS content analysis is performed using text pre-processing and analyzing techniques to detect the presence of URL, Phone Number, E-mail ID, and malicious keywords in the message. We have used a machine learning algorithm to classify the message on basis of malicious keywords present in the message. We have also used the techniques of form tag check and APK download check to analyze the malicious behavior of the URL. Text messages will be finally classified into a malicious and non-malicious category based on the results of the detection techniques.

Keywords: Smishing, Phishing, Text Messaging, Machine Learning, Mobile Security, SMS Short Message Service