Malware behavior feature extraction based on web information extraction

Malware is one of the major threats on the Internet. This paper presents a new method which extract malware behavior feature from the analysis report on the web. With the malware behavior information from the report, we construct the malware behavior feature library. The extraction process is automatic and the extraction results can be read directly by user. Through mining the behavior feature library, we could get the characteristics of malware. The evaluation shows the method with good precision rate and recall rate.