Open Access Open Access  Restricted Access Subscription or Fee Access

Comprehensive Evolution of Different Methods Used in Data Mining-Based Intrusion Detection System

Chintan M. Bhatt, Amit P. Ganatra, C.K. Bhensdadia, Yogeshwar P. Kosta

Abstract


Intrusion is defined as an invasion that consists of set-of-actions that compromise upon the integrity, confidentiality or availability of data-resource/s. Therefore, intrusion detection is an important task when dealing with an information infrastructure for security. A major challenge in intrusion detection is to unearth intrusions that happen almost instantaneously and thereafter lay embedded, to be discovered, in vast scattered resources in a normally operating real-time communication environment. Data mining process working on intrusion detection is to identify valid, novel, potentially useful, and ultimately understandable patterns in massive data. Thus, it can be understood that, it is challenging as well as demanding to apply data mining techniques to detect intrusions of various types in an information infrastructure resource/s. To start with, our paper discusses different intrusion detection techniques that brings out and presents the underlying concepts and associated application of data mining approaches as an applied tool against intrusion detection system. Techniques include, Support Vector Machines (SVMs) that was designed and utilized as classifiers for binary classification/s, and helped to solve multi-class problems. In this paper we bring in the fusion of Decision-Tree and Support Vector Machine (DT-SVM) which combines and reinforce in an effective way for solving multi-class problems in the information resource domain. This method has the potential, as confirmed in our findings, to decrease the training and testing time, contributing to increased efficiency of the system. The construction order of binary tree significantly influences classification performance. Towards the end of the paper we report aspects relating to development of an algorithm that combines to produce a Tree structured multi-class SVM as an intrusion detection data mining technique, which has been applied successfully for the purpose of classifying data that aid the process of intrusion detection.

Keywords


Ant-Miner, COD (Common Outlier Detection), Decision Tree, Fuzzy C-Means, K-Means, MACO, Support Vector Machine (SVM) and Decision-Tree and Support Vector Machine (DT-SVM).

Full Text:

PDF

References


Bo Liu , Hussein A Abbass and Bob Mackay, “Classification Rule Discovery with Ant Colony Optimization”, IEEE/WIC International Conference on Intelligent Agent Technology (IAT’03)

Junbing He, Dongyang Long and Chuan Chen, “An Improved Ant-based Classifier for Intrusion Detection”, IEEE Third International Conference on Natural Computation (ICNC 2007).

Junzhong Ji, Ning Zhang and Chunnian Liu, “An Ant Colony Optimization Algorithm for Learning Classification Rules”, IEEE/WIC/ACM International Conference on WebIntelligence (WI06).

[4] Zhengxim Chen, “Data Mining and Uncertain Reasoning-An integrated approach”, Willey, 2001.

[5] J.Bezkek, “pattern Recognition with fuzzy objective function algorithms”, Plennum Press, USA, 1981.

[6] S.Albayrak, and Fatih Amasyali, “Fuzzy CMeans clustering on medical diagnostic systems”, International XII Turkish Symposium on Artificial Intelligence and Neural Networks, TAINN-2003.

[7] Wit old Pedrycz, “Knowledge Based Clustering”, John Willey&sons Inc., 2005.ISBN:0-471-46966-1.

[8] Witcha Chimphlee, et.al. “Un-supervised clustering methods for identifying Rare Events in Anomaly detection”, in Proc. Of World Academy of Science, Engg. and Tech (PWASET), Vol.8, Oct2005, pp.253-258.

[9] Rafael S Parpinelli, Heitor S Lopes and Alex A Freitas, “Data Mining With and Ant Colony Optimization Algorithm”, IEEE Transaction on Evolutionary Computing, Vol. 6, No. 4, August 2002.

[10] Quinlan JR. C4.5: programs for machine learning. Log Altos,CA: Morgan Kaufmann; 1993.


Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.