Effective classification of android malware families through dynamic features and neural networks

Due to their open nature and popularity, Android-based devices have attracted several end-users around the World are one of main targets for attackers. Because reasons given above, it is necessary build tools that can reliably detect zero-day malware on these devices. At moment, many frameworks been proposed applications leverage Machine Learning (ML) techniques. However, an essential requirement consists using very large sophisticated datasets model construction training purposes. Their success, indeed, strongly depends choice right features used building a classification providing adequate generalisation capability. Furthermore, creation dataset well represents properties behaviour most critical challenges in analysis. Therefore, aim this paper proposing new called Unisa Malware Dataset (UMD) available http://antlab.di.unisa.it/malware/, which based extraction static dynamic characterising activities. Additionally, we will show some experiments concerning common ML demonstrate how possible efficient ML-based dataset.