Blockchain for decentralized malware detection on android devices

Abstract

The increasing use of Android smartphones has led to an increase in malware threats, which calls for quick fixes that standard antivirus software cannot provide. In light of this, the paper investigates the revolutionary idea of "Decentralized Malware Detection for Android Using Blockchain." This innovative method offers a decentralized, highly flexible malware detection system using blockchain technology, which was initially created for cryptocurrencies. This approach aims to drastically lower the likelihood of false positives and false negatives, in contrast to existing security solutions, which are notorious for their reliance on frequent updates and susceptibility to evasion techniques. To do this, the malware detection process is distributed among several distinct nodes, each of which is responsible for independently verifying the authenticity of an Android application. In the proposed work we have utilized the dataset of Androzoo containing the data for more than five million apps. Based on the Virus total VT score we segregated this data and created our own dataset containing the permission-based & network service-based features. We have applied the machine learning techniques like Random Forest, Decision tree, Adaboost & K-NN whose performance accuracy has been utilized by the decentralized blockchain node for effective malware detection. We have achieved the malware detection accuracy of around 90% when sufficient features of Android applications available for analysis. The results presented in the proposed work aims to present a comprehensive understanding of how blockchain technology can be applied to enhance the detection of Android malware, thereby improving mobile apps security and resilience in the mobile ecosystem