Detecting Malicious PDF Document Using Supervised LearningAlgorithm

In this paper we investigate the vulnerabilities of Portable Document Format, which cause client side attack and introduce a detection system with capable of detecting malicious PDF documents that are transferred over a network. Proposed detection system improves the rate and accuracy of previous way and performs classification using a machine learning classifier which tested with created new dataset. We designed a tool for immigrate from object level to code level and we designed three feature groups that were relevant and significant for the classification of PDF documents as benign or malicious. Then we classified with machine learning algorithm and compared each other. Identify significant feature for detecting malicious PDF files in supervised learning algorithm is major achievement in this paper. Also building PDF dataset and ensure that all factors are under control is another result of this research.

Detecting Malicious PDF Document Using Supervised LearningAlgorithm Keywords:

Machine learning , Malicious PDF , Detection system , Classification , Supervised algorithm