Opcode based malware detection using Stacked Recurrent Neural Networks and Convolutional Neural Networks

Malwares is designed to infiltrate devices and in most cases the underlying operating system and are packaged in multiple file formats. Considering that Windows operating system is one of the most opted operating systems worldwide in this paper we are focusing on developing a PE file malware detection system using Stacked CNN and RNN. In this paper we are training a LSTM and GRU model along a CNN model to detect malware PE files. We are using two techniques for comparing the results, we are using the DikeDataset which is an open source active dataset for collecting malware and benign files. Further we are employing static analysis using the opcode frequency based malware detection technique. Further we are exploring the Blue Hexagon Open Dataset for Malware AnalysiS (BODMAS) dataset which is a comprehensive dataset maintained for 57,293 malware samples and 77,142 benign samples and 581 malware families and the EMBER dataset.