Android traffic malware analysis and detection using sprint machine learning technique

Abstract

The recent history has seen cybercrime seize the notorious position of stealing business and personal information of individuals. This risk is because of vulnerability in security defense and the explosion in the use of interconnected devices. Malware is created in various types; they include viruses, worms, trojans, ransomware, spyware, and adware. Among all the malware, the Android malware is the worst because it can destroy the device by invading the privacy of a person and stealing his/her information. The already adopted anti-virus and anti-malware software require periodic updates and have no ability to detect the new and advanced malware. Our paper, therefore, proposes the malware analysis and systematic classification using machine learning algorithms. In our study, we associate the effectiveness of Deep Neural Network, Recurrent Neural Network, and Secure Program Identification  classifiers with the features of Ack-post, Finance, Moghave, Send pay and Chess to pre-train and test the models. We have obtained a classification accuracy of 99.5% with the proposed SPRINT classifier, 98 percent with DNN, and 96 percent with RNN, through our experiment. Such an outcome highlights the superiority of the proposed SPRINT classifier to the rest in its ability to classify the malware with machine learning algorithms. With the help of the proposed approach, new and advanced malware instances can be timely and effectively identified and classified, which would contribute to the improved security of Android devices and eliminate possible cyber threats.