Detection of android malwares using portable executable files with LSTM model

In computer programming, malicious code refers to programs that are intended to do damage to a system by introducing or exploiting flaws. In turn, this may lead to security lapses, data breaches, and other types of file and system harm. Malicious software threats and their detection are becoming more important as a sub domain of Information security. There have been studies that are recently made on malware detection approaches. In the design and development of anti-malware systems, the most difficult challenge is ensuring that all malware is detected. Deep Learning was utilized to increase the identification of malware variants and applying a structure for identifying the vicious code by examining trace files using Long Short-Term Memory. We modeled the run traces of the malicious and non-malicious PE files. For the initial dataset,PE files are obtained by Dynamic Analysis, which we then used to test our hypotheses on a real-world dataset comprised of both helpful and harmful applications. Our model was shown to be accurate and fast by the results of the experiments.