Year: 2013
Type: dissertation
Abstract: ENGLISH: The PDF-file format is a very popular format to perform attacks with due to the format being quite\nversatile. A PDF-file can be used as direct attacks against specific targets like the government,\nthe army or other high value targets. These kinds of attacks may be performed by foreign\nintelligence or by organised crime because they have the most to gain by a successful attack.\nThe attacks are often well obfuscated which makes it easy for users to unintentionally execute\nthe malware on his/her machine. A PDF-file may for instance contain a well written report with\nimportant information to the user [1], but do also contain malicious code in order to perform\nreconnaissance on the target’s network.\nThis master thesis is a continuation of the results of Jarle Kittilisen’s master thesis in 2011. The\nthesis will utilize Kittilsen’s proposed methodology by using the machine learning tool ’support\nvector machine’ in order to classify PDF-files as malicious or benign. This thesis will focus on\nonline detection of PDF-files where as Kittilsen performed post-detection. One of the biggest\nproblems with an online detection of PDF-files is the time frame from the PDF-file is detected\nuntil it has been classified as either malicious or benign. This master thesis seek to provide\nanswers for the viability of an online detection system of PDF-files.
Author: Knut Borg
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2464596756