抽象的な

A Survey of Intrusion Detection System Using Different Data Mining Techniques

Trupti Phutane, Apashabi Pathan

Now- a- day authentication is of prime concern. We have to look every aspects of security in every transaction to avoid the treats & intruders. In this paper, we will discuss the existing intrusion detection systems with data mining approach such as intrusion detection system using association rule mining [6] and intrusion detection system using event correlation data mining [5]. We also discuss the proposed system [3], [4], [10] in which the intrusion detection is a data analysis process, rather than previous approaches like knowledge engineering processes. In association rule mining, we firstly capture the network data using sniffers. The captured data will be filtered so that non relevant data is removed from analysis and finally we will extract the features which will be associated with given datasets where as in event correlation data mining method we will maintain the logs of every network system and on the basis of “event data” stored in logs we will try to maintain the link among them, if we found any suspected activity we will deny the access for data. In both the techniques we are only concerned with the allowing & denial of access but in proposed system we are maintaining the decision tree by analysis of data and its attributes rather than just guessing or finding any relations with previous data.