SEDMDROID An Enhanced Stacking Ensemblr Framework For Android Malware Detection

Suggest a architecture that enables the use of various machine learning algorithms to effectively differentiate between malware files and clean files while attempting to reduce the amount of false positives. In this research, we first work with cascade one-sided perceptrons and then with cascade kernelized one-sided perceptrons to demonstrate the concepts underlying our system. The concepts underpinning this framework were tested successfully on medium-sized datasets of malicious and clean files before being put to a scaling-up Process that enables us to work with very big datasets of malicious and clean files.