global-journal-of-computer-science-and-technology-c-software-data-engineering Global Journal of Computer Science and Technology - C: Software & Data Engineering 0975-4350 0975-4172 Global Journals Publishing Group Incorporated 77872 Optimized Anomaly based Risk Reduction using PCA based Genetic Classifier C.Kavitha K.IyakuttiDr. INDIA, Pasumpon Muthuramalinga Thevar College,Usilampatti, Madurai 15 01 2014 14 C7 31 37

Security risk analysis is the thrust area for the information based world. The researchers in this field deployed numerous techniques to overcome the information security oriented problem. In this paper the researcher tried for a approach of using anomaly detection for the risk reduction. The hub initiative for this work is that the anomalies are the deviation which could increase the percentage of risk. The anomaly detection is guided by the PCA and the genetic based multi class classifier is used. The classification is induced by the decision tree approach were the genetic algorithm is set out for the optimization in the process of finding the nodes of the tree. The proposed approach is evaluated with the bench mark on PCA based ANN classifier. The proposed approach outperforms the existing one. The results are demonstrated.

 anomaly detection PCA genetic algorithm. Full Text

Security risk analysis is the thrust area for the information based world. The researchers in this field deployed numerous techniques to overcome the information security oriented problem. In this paper the researcher tried for a approach of using anomaly detection for the risk reduction. The hub initiative for this work is that the anomalies are the deviation which could increase the percentage of risk. The anomaly detection is guided by the PCA and the genetic based multi class classifier is used. The classification is induced by the decision tree approach were the genetic algorithm is set out for the optimization in the process of finding the nodes of the tree. The proposed approach is evaluated with the bench mark on PCA based ANN classifier. The proposed approach outperforms the existing one. The results are demonstrated.