Informatics Publishing Limited

P. Prasenna ¹, A. V. T. Raghav Ramana ²

Affiliations
1 Oxford Engineering College, Tiruchirappalli, Tamil Nadu, IN
2 Department of Information Technology, Oxford Engineering College, Tiruchirappalli, Tamil Nadu, IN

Abstract

In conventional network security relies on mathematics cryptosystem and low counter security measures to taken to prevent Intrusion detection System, although most of this approaches in terms of theoretically impossible to implement. One of the Evolutionary optimization techniques like Genetic Network Programming (GNP) is node based directed graph structures instead of generating a large number of rules and patterns, In this paper focusing on generalize the problem embedded in GNP with association rule mining and address to a issues in IDS and gives a solution to detecting intrusion. Our proposed method follows an Apriori algorithm based fuzzy WAR and GNP and avoids pre and post processing thus eliminating the extra steps during rules generation. This method can sufficient to evaluate misuse and anomaly detection. Experiments on KDD99Cup and DARPA98 data show the high detection rate and accuracy compared with other conventional method.

Keywords

Intrusion Detection, Probability Density Function, Genetic Network Programming, Genetic Algorithm, Fuzzy WAR.

Full Text

     

R. Hariharan ¹, C. Mahesh ¹, P. Prasenna ¹, R. Vinoth Kumar ¹

Affiliations
1 Department of Information Technology, Vel Tech University, Chennai - 600062, Tamil Nadu, IN

Abstract

Background/Objectives: Privacy preservation in data mining to hold back the sensitive data from attackers. Findings: There are various existing methods available to preserve the data like perturbation, anonymization, randomization etc., each method has its own advantages and disadvantages. The trade-off between security and utility of data should be handled with standardizing methods for the PPDM. In this paper explained a method based on PPDM in data mining using cluster based greedy method. Application/Improvements: This method can be applied in sensitive data areas such as hospitals, Customer Management System, government survey, etc., where there is need for privacy preservation.

Keywords

Cluster based Greedy Method, Classification Error, Isometric Transformation, Privacy Preservation, Privacy Preservation Rate

Full Text