Behaviour profiling on mobile devices

Over the last decade, the mobile device has become a ubiquitous tool within everyday life. Unfortunately, whilst the popularity of mobile devices has increased, a corresponding increase can also be identified in the threats being targeted towards these devices. Security countermeasures such as AV and firewalls are being deployed; however, the increasing sophistication of the attacks requires additional measures to be taken. This paper proposes a novel behaviour-based profiling technique that is able to build upon the weaknesses of current systems by developing a comprehensive multilevel approach to profiling. In support of this model, a series of experiments have been designed to look at profiling calling, device usage and Bluetooth network scanning. Using neural networks, experimental results for the aforementioned activities' are able to achieve an EER (Equal Error Rate) of: 13.5%, 35.1% and 35.7%.