Android application vulnerabilities are weaknesses in the applications or their design that allow an attacker to execute commands, access unauthorised data and carry out denial of service attacks. In this paper, we try to improve the detection of vulnerabilities, especially those based on permissions in Android applications. We propose a model that uses machine learning techniques for the automatic detection of vulnerabilities in the permissions system on Android devices. This model allows the user to identify the risk of vulnerabilities associated with an application via a permission or combination of permissions. We applied eight (8) machine learning techniques to analyse the use of permissions for malware detection. Our goal is to enable users to assess the vulnerabilities of applications that may access their personal and other sensitive data.

Vulnerability, Permission-based, Machine learning, Android applications