In this paper is presented an approach for behavioral prediction of network attacks usinghoneypot rules using support physical machine as a classification technique. Theproposed honeypot is used to trap delay and gather information about the attackers.These systems are centralized at a strategic point in the network and collect data into alog file that is subsequently analyzed as main vulnerability. Honeypots can generate thesignature with the help of attack classifier, and it can detect and record known andunknown attacks in addition. Signature-based classifier uses Bayesian classificationalgorithm to represent malicious file or normal file. Therefore, the IP addressescorrespond to higher chances for detecting intruders as more fake systems would bemined in the network.Keywords: Honeypot, attacks defined, signature generation, malicious, nonmalicious,IP blacklisting
