Abstract: The CVE dictionary can not design effective defense strategies for clustered system’s vulnerabilities because of it’s lack of categorization and generalization capabilities.To address such shortcomings,this paper propose a CVE categorization framework called CVE Classifier,which can categorize CVEs according to various taxonomic features.CVE classifier automatically extract the training data from the associated vulnerability databases (including:BID,X-Force and Secunia) and builds learning models for taxonomic features based on it by using SVM.

Key words: Support Vector Machine(SVM), Common Vulnerabilities and Exposures(CVE), taxonomic feature