A TWO-DIMENSIONAL DATA FUSION MODEL FOR INTRUSION DETECTION

摘要：

When the same data are detected and classified with different classifiers, there will be inconsistencies in the results. This shows that different factors cause the classifiers detection accuracy not alike. In this study, the proposed methods were verified with KDOCUP99 data, and data fusion (DK) using five feature selection methods (Discriminant Analysis, DA; Principal Component Analysis, PCA; Rough Set Theory, RST; Multiple Logistic Regression, MLR and Genetic Analysis, GA.). In the case of data re-determination and upgrading the detection was accurate. In this study, we propose two dimensional DF. Combining different DF methods can increase the IDS detection accuracy. Empirical results using a KDDCUP99 dataset had an intrusion detection accuracy of 99.9834％, which made it useful for intrusion detection and data re-determination.

关键词： Intrusion detection system Data fusion Dempster-Shafers Theory Bayesian Theory Support Vector Machine

作者: KUN-MING YU MING-FENG WU

作者单位: Department of Computer Science and Information Engineering, Chung Hua University, No.707, Sec.2, WuF Department of Information Management, Chung Hua University Chung Hua University, No.707, Sec.2, WuFu

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 3970-3974

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）