Permission based Android Malicious Application Detection using Machine Learning
Aditya Kapoor, Himanshu Kumar Kushwaha, Ekta Gandotra · 2019
Since the launch of the smartphones, their usage is increasing exponentially. They have become an important part of our lives. We are very much dependent on smartphones for our daily routine and use numerous applications both from the play store or the third party applications. Most of the times, the applications downloaded from unofficial sources pose a threat as there doesn't exist the necessary checks or mechanisms to validate the authenticity of these applications and maybe infected with malware. The malware infected applications can lead to leakage of user's personal data. Anti-virus tools use signature based methods for detecting malwares, but their databases need to be updated regularly. In this paper, we present a system for classifying Android applications on the basis of permissions used by those applications. We used six machine learning algorithms for classifying these applications into malicious or benign applications. On comparing the results, we find that Logistic Regression Algorithm suits best to our dataset and provide 99.34% accuracy.