Performance Enhancement of Intrusion Detection System and Scheme of Data-Dependent Decision Fusion System | Original Article

ABSTRACT:

Rapid development in networking technology contributes to an unprecedented growth in the amount of unwanted or harmful network acts. As a defense-in-depth portion, the Network Intrusion Detection System (NIDS) is designed to identify malicious behaviours. Usually, NIDSs are being applied using a variety of classification techniques, but some techniques are not developed sufficiently detect complex or synthetic attacks accurately, particularly when confronted with large, high-dimensional data. In addition, the intrinsic drawbacks of NIDSs, including high false alarm rate and poor detection performance, have not been successfully resolved. To order to address these issues, data fusion (DF) was extended to the attack prevention of the network and has obtained successful performance. Nevertheless, there is also a shortage of in-depth research and application of data fusion technologies in the area of intrusion detection. This is also important to carry out a thorough analysis of these steps. In this paper, we concentrate on DF strategies for the identification of network intrusions and suggest a common description to explain it. We study the recent developments in DF techniques and suggest a set of parameters for evaluating their efficiency. Subsequently, based on the findings of the literature review, a range of accessible problems and potential strategies for study are suggested at the conclusion of this article.