摘要
建立在统计学习理论和结构风险最小原则上的支持向量机(SVM)在理论上保证了模型的最大泛化能力,因此将支持向量机理论应用于入侵检测领域可以获得很好的效果。但是在应用中也存在如何对网络数据进行特征编码和选择适当的支持向量机模型参数的问题。在分析了特征编码和模型参数对分类器识别精度的影响基础上,提出用遗传算法建立支持向量机带权特征和分类器模型参数的自适应优化算法,并在网络入侵检测中成功的运用算法。最后,使用KDD CUP 1999数据进行的仿真实验表明了算法的正确有效性。
The support vector machine（SVM）,which is based on the statistical learning theory and the structural risk minimum principle,guarantees the largest generalization ability of a model.So it can obtain very excellent effect when using the SVM theory in the field of intrusion detection.However,in the application it also has some problems such as how to code the features of network data and how to select the proper model parameters of SVM.The paper proposed a self-adaptive optimization algorithm for the SVM feature with weight and model parameters using genetic algorithm on the basis of analyzing the influence of feature coding and model＇s parameters on the classifier＇s recognition accuracy.And the algorithm was successfully applied for network intrusion detection.Finally,experiment shows the effectiveness of the optimization algorithm using the KDD CUP 1999 Data.
出处
《计算机仿真》
CSCD
2008年第9期115-117,158,共4页
Computer Simulation
基金
浙江省自然科学基金(Y106176)
浙江省科技厅科技计划项目(2007C33058)
关键词
遗传算法
支持向量机
入侵检测
Genetic algorithm
Support vector machine（SVM）
Intrusion detection
