Abstract

Malicious software continues to pose a major threat to the cyber world. Text files are the most frequently used vectors to infect various systems using malware. In all this, to execute the attack, the intruder attempts to merge the malignant code with the benevolent text data. Due to its compatibility and lightweight characteristics, PDF (portable document format) is the most widely used file method of sharing documents. In today's world, attackers are using cutting-edge methods to obfuscate malware concealed inside document files. So, it is difficult for malware detection classifiers to effectively identify the text. To understand their design and working procedures, we surveyed different types of learning-based PDF malware classifiers. Also, we have described the pdf document by which we can understand the workings of malware. Finally, we recommended a methodology on the basis of the literature survey and specified the future direction for the better classification results. This work is the extension of dissertation.