A Malware Detection Scheme Based on Mining Format Information
Authors:
Open Access
- 2 June 2014
- journal article
- research article
- Published by Hindawi Limited in The Scientific World Journal
- Vol. 2014, 1-11
Abstract
Malware has become one of the most serious threats to computer information system and the current malware detection technology still has very significant limitations. In this paper, we proposed a malware detection approach by mining format information of PE (portable executable) files. Based on in-depth analysis of the static format information of the PE files, we extracted 197 features from format information of PE files and applied feature selection methods to reduce the dimensionality of the features and achieve acceptable high performance. When the selected features were trained using classification algorithms, the results of our experiments indicate that the accuracy of the top classification algorithm is 99.1% and the value of the AUC is 0.998. We designed three experiments to evaluate the performance of our detection scheme and the ability of detecting unknown and new malware. Although the experimental results of identifying new malware are not perfect, our method is still able to identify 97.6% of new malware with 1.3% false positive rates.Keywords
This publication has 14 references indexed in Scilit:
- Detecting unknown malicious code by applying classification techniques on OpCode patternsSecurity Informatics, 2012
- Monitoring, analysis, and filtering system for purifying network traffic of known and unknown malicious contentSecurity and Communication Networks, 2011
- Unknown malcode detection and the imbalance problemJournal of Computer Virology and Hacking Techniques, 2009
- Detection of malicious code by applying machine learning classifiers on static features: A state-of-the-art surveyInformation Security Technical Report, 2009
- PE-Miner: Mining Structural Information to Detect Malicious Executables in Realtime2009
- Detection of unknown computer worms based on behavioral classification of the hostComputational Statistics & Data Analysis, 2008
- Unknown Malcode Detection Using OPCODE Representation2008
- Reliable identification of bounded-length viruses is NP-completeIEEE Transactions on Information Theory, 2003
- A Decision-Theoretic Generalization of On-Line Learning and an Application to BoostingJournal of Computer and System Sciences, 1997
- Smoking out the bad actors: Risk analysis in the age of the microcomputerComputers & Security, 1989