Adaptive Intrusion Detection Algorithm Based on New Conditional Entropy
Qian Luo · Computer Technology and Development · 2011
Based on the analysis of the current intrusion detection approaches,existing security detection systems have many problems such as wrong detection of intrusions,missed intrusions,poor real-time performance,bring up a new detection method,namely adaptive intrusion detection algorithm based on new conditional entropy.In considering the theories related to information theory,this algorithm firstly discrete the collected data use the knowledge of information entropy,then analyze the discrete data,remove the redundant attributes by reduction method related to conditional entropy knowledge,finally generate a new detection rules for the further analysis of intrusion data.The experimental result shows that is more efficient than algorithms based on BP neural networks and vector machines;thereby,this detection algorithm can effectively improve the intrusion detection system's detection rate,and reduce the error detection rate,and this detection algorithm can improve the detection ratio by about 7% and reduce the wrong detection ratio.The system provides detection service effective for information systems,as well.