The malware threats for mobile phones continues to rise as demand has prompted the development process to mainly focus on adding new attractive features. Unfortunately, this exponential growth of mobile devices is not keeping pace with design of new security solutions before these threats can inflict widespread damage. Many business systems and networks are the main victims of malicious attacks by worms, viruses, spyware and other intrusion activities to cripple even the most critical success services. There are reports suggesting that combining spyware as a malicious payload with worms as a delivery agent has generated malicious programs that can be used for industrial espionage and identity theft. In this paper we propose a new behaviour approach using machine learning algorithms for detecting existing and emerging malware targeting mobile phones. The approach is basically focusing on the concept of a generalised behaviour pattern with additional emphasis on detecting new classes of malware that integrate attributes or features from existing classes of malicious malware bodies. The evaluation experiments demonstrate that different levels of abnormal behaviour were accurately detected.

Keywords: malicious malware, SVM, mobile malware, Behaviour-based malware, machine learning.

.