Selecting Features to Classify Malware

Download Now
Provided by: Adobe Systems Incorporated
Topic: Security
Format: PDF
Malicious computer programs, or malware, are a pressing business and personal computing problem. Malware evolves rapidly. Detecting new malware, which may just be variants of existing malware, is a strain on antivirus engines. Malware is a menace to computing. The lag between malware landing on a user's system and the development of signatures to detect the same malware can prove catastrophic for users. Using data mining, the authors identify seven key features within the Microsoft PE file format that can be fed to machine learning algorithms to classify malware. The resulting models classify malware with results comparable to existing research that uses many more features.
Download Now

Find By Topic