Entity

Time filter

Source Type

Gurgaon, India

Bahl S.,KIIT University | Sharma S.K.,Ansal University
Advances in Intelligent Systems and Computing | Year: 2015

Intrusion detection system (IDS) research field has grown tremendously over the past decade. Most of the IDSs employ almost all data features for detection of intrusions. It has been observed that some of the features might not be relevant or did not improve the performance of the system significantly. Objective of this proposed work is to select a minimal subset of most relevant features for designing IDS. A minimal subset of features is chosen from the features commonly selected by correlation based feature selection with six search methods. Further, the performance comparison among seven selected subsets and complete set of features is analyzed. The simulation results show better performance using the proposed subset having only 12 features in comparison to others. © Springer International Publishing Switzerland 2015. Source


Bahl S.,KIIT University | Sharma S.K.,Ansal University
International Conference on Advanced Computing and Communication Technologies, ACCT | Year: 2015

Intrusion detection system (IDS) research field has grown tremendously in the past decade. Improving the detection rate of user to root (U2R) attack class is an open research problem. Current IDS uses all data features to detect intrusions. Some of the features may be redundant to the detection process. The purpose of this empirical study is to identify the important features to improve the detection rate and reduce the false detection rate. The investigated feature subset selection techniques improve the overall accuracy, detection rate of U2R attack class and also reduce the computational cost. The empirical results have shown a noticeable improvement in detection rate of U2R attack class with feature subset selection techniques. © 2015 IEEE. Source


Borkar M.,Ansal University | Nitin,Jaypee Institute of Information Technology
Journal of Supercomputing | Year: 2015

This paper proposes a new variant of Gamma Interconnection Network (GIN), which uses an alternate source at the initial stage. The alternate source helped in realizing the Bit Reversal permutation completely in one pass. The paper also proposes a modified permutation realization algorithm, which is being used to realize the frequently used permutations on GIN family of networks. This algorithm also ensures that the alternate source approach can be used with all the GIN family networks with sizes $$\ge $$≥8 to realize the frequently used permutations. The paper also discusses the performance of the modified algorithm in terms of hop count required to realize the permutations as well as the effect on hardware cost due to alternate source. © 2015, Springer Science+Business Media New York. Source


Aggarwal P.,Ansal University | Sharma S.K.,Ansal University
Advances in Intelligent Systems and Computing | Year: 2015

Intrusion Detection System (IDS) can be called efficient when maximum intrusion attacks are detected with minimum false alarm rate but due to imbalanced data, these two metrics are not comparable on the same scale. In this paper, a new NPR metric is suggested in view of the imbalanced data set to rank the classification algorithms for IDS which can help analyze and identify the best possible combination of high detection rate and low false alarm rate with maximum accuracy and F-score. The new NPR metric is used for comparison and ordering of ten classifiers simulated on KDD data set. © Springer International Publishing Switzerland 2015. Source


Bahl S.,Ansal University | Sharma S.K.,Ansal University
International Conference on Computing, Communication and Automation, ICCCA 2015 | Year: 2015

Intrusion detection system (IDS) research field has grown tremendously in the past decade. Improving the detection rate of user to root (U2R) attack classes is an open research problem. Current IDS uses all data features to detect intrusions. Some of the features may be redundant to the detection process. The purpose of this empirical study is to identify important features to improve the detection rate of U2R attack class. The investigated correlation feature selection improved the overall accuracy, detection rate of U2R attack. The empirical results have given a noticeable improvement in detection rate of U2R. © 2015 IEEE. Source

Discover hidden collaborations