A Review on High Ranked Features based NIDS

Abstract: With the rapid growth in the network traffic day by day the new threats are evolved affecting network security. The benchmark KDD dataset which was generated a decade ago has become outdated as it does not inclusively reflect modern normal behaviors and contemporary synthesized attack activities. In this paper we have used a new UNSW-NB15 data set and compared with the KDD data set and its version. As the network packets consist of a wide variety of features containing some irrelevant and redundant features which reduces the efficiency of detecting attacks, and increase False Alarm Rate (FAR). So to choose the relevant features and remove the redundancy we used central points of attribute values and association rule mining algorithms which help in reducing the processing time by selecting the most frequent values. These algorithms are applied on KDD99 and UNSW-NB15 data sets to get the high rank features.

Keywords: UNSW-NB15 and KDD99 data set, Central point, association rule mining, features Selection.