Remote Operation System Detection Base on Machine Learning

A machine learning method for remote operation system recognition through their detection signatures with support vector machine (SNM) is proposed. A vector space model of Nmap fingerprint database and techniques for translating the host responses to SVM input vectors are also suggested. Experimental result on identification of signatures in the fingerprint database of Nmap 4.90RC1 but not known for Nmap 4.76 show that our method is effective in the discovery of new signatures not included in current fingerprint database.