Intelligent Optimal Gated Recurrent Unit based Malicious PDF Detection and Classification Model

Because of the familiarity of the portable document format (PDF) and raising number of susceptibilities in most of the PDF viewer applications, malware writers endure to utilize it to distribute malware through web downloads, email attachments, etc in targeted as well as non-targeted attacks. Malicious PDF files signify a major threat to computer security and several research works have started to resolve the issue with ensuring security. Numerous approaches are designed to examine and identify the existence of the malicious content in the PDF, but the basic idea to design malicious PDF detection approaches has not been completely considered. In this aspect, this paper designs an intelligent deep learning based malicious PDF detection and classification (IDL-MPDF) model. The IDL-MPDF technique involves a feature engineering process using the PeePDF tool, which scans the structure of the PDFs, returns the tags and keywords which are utilized to create them. Besides, gated recurrent unit (GRU) model is employed for the malware PDF classification and the hyperparameter optimization of the GRU model takes place using the whale optimization algorithm (WOA). In order to investigate the goodness of the IDL-MPDF technique, an extensive simulation analysis is carried out and the experimental results highlighted the promising performance interms of different measures.