Most intrusions on our devices and computers through malware, third-party attacks, and other infractions tend to go unreported unless one is particularly tuned into the information security world and knows the procedures needed to get the message to the right people. Research published in the International Journal of Internet Technology and Secured Transactions, looks at how machine learning might be used to automate the process of reporting policy violations on a system.
Albara Awajan, Moutaz Alazab, Issa Qiqieh, and Mohammad Wedyan of Al-Balqa Applied University in Al-Salt, and Salah Alhyari of JEPCO in Amman, Jordan, point out that computer and mobile devices users frequently face security incidents and violations of their systems and data. They point out that a unified approach to reporting such malicious activity could, to some degree, address this growing problem. They have now proposed an automated client-server citizen reporting system framework based on machine learning techniques that could help.
The system can classify images a user wishes to use to accompany a report and can be used to report any cyber-crime incidents such as bank account intrusion, credit card fraud as well as phishing and pharming attacks on their devices. Tests demonstrated that the new framework is fast, convenient, and performs effectively and efficiently with different mobile devices using the common Android operating system. Classification accuracy is 95.4% and a prediction time of just 5.30 seconds.
The team is now optimising the framework as well as investigating how it might be extended to other additional smartphone operating systems such as Apple iOS, Windows Phone, and the Huawei operating system.
Awajan, A., Alazab, M., Alhyari, S., Qiqieh, I. and Wedyan, M. (2022) ‘Machine learning techniques for automated policy violation reporting’, Int. J. Internet Technology and Secured Transactions, Vol. 12, No. 5, pp.387–405.