Android malicious application detection using permission vector and network traffic analysis

In this technology world, smartphones are greatly adopted by people due to the need of personal communication, Internet and many more requirements. Users are attracted to use the android operating system due its availability for low-cost and millions of freely available applications. The popularity of android operating system is also welcomes the attackers. Statistics have shown that, the growth of android malware is becomes double by every year. Hence android platform is more vulnerable to malwares. Researchers are proposed various models. Some of these models are completely fail to detect unseen variants of malware, while remaining models are inefficient to detect new malware families. In this paper, we briefly explain about android architecture, structure of android application and also characterized android malware based on their installation, activation and payloads types. We proposed a hybrid model to detect the malware based on permission bit-vector and network traffic. We constructed a decision tree classifier to detect the android malware. Our results show that combination of permission bit-vector and network traffic analysis is highly efficient by achieved 95.56% of detection accuracy.