摘要
针对支持向量机增量算法中边界样本的提取机制效率不高的问题,提出基于云模型的增量SVM入侵检测方法。该算法利用云模型稳定性和不确定性的特点,将异类样本间的特征距离映射成隶属度函数,对初始集中边界向量进行提取。分析新增样本对支持向量集的影响,淘汰无用样本。理论分析和仿真实验表明,该算法在保证分类精度的同时有效地提高了检测速度。
A new incremental SVM intrusion detection algorithm based on cloud model is proposed for the problem of SVM incremental algorithm that the boundary samples extraction mechanism is of low efficiency.In this algorithm,the stability and uncertainty characteristics of the cloud model is used to map the feature distance between the heterogeneous samples onto the membership function to extract the boundary vectors of the initial dataset.At the same time the impact of the new samples added on the support vector set is analysed,and the useless samples are discarded.Theoretical analysis and simulation experiments demonstrate that the detection speed is greatly improved while the classification precision is guaranteed.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第3期311-314,共4页
Computer Applications and Software
关键词
入侵检测
支持向量机
增量学习
云模型
边界向量
Intrusion detection Support vector machine（SVM） Incremental learning Cloud model Boundary vectors
