蚁群算法选择特征与WSVM融合的网络入侵检测  

Network intrusion detection with ant colony optimization feature selection and WSVM

作　　者：冯庆华[1] 

机构地区：[1]江苏建筑职业技术学院实验实训与职业技能管理中心,江苏徐州221116

出　　处：《江苏建筑职业技术学院学报》2014年第3期38-41,共4页Journal Of Jiangsu Vocational Institute of Architectural Technology

摘　　要：为了提高网络入侵检测率,提出一种蚁群算法选择特征与加权支持向量机的网络入侵检测方法.利用蚁群算法选择网络数据的关键特征,计算信息增益获得各个特征权重,根据特征权重构建了加权支持向量机的网络入侵分类器,并通过KDD CUP 99数据集验证了其有效性.结果表明:该算法能够有效降低特征维数,提高网络入侵检测率和检测效率.In order to improve the detection rate of network intrusion,this paper proposes a net-work intrusion detection method with ant colony optimization feature selection and weighted sup-port vector machin.By selecting the key features of network data with ant colony optimization, we calculates information gain to get each feature weight,establishes network intrusion classifier of weighted support vector machine according to feature weight,and verifies the validity through KDD CUP 99 dataset.Results show that ACO-WSVM can reduce the feature dimension effec-tively and improve network intrusion detection rate and efficiency.

关 键 词：网络入侵检测 蚁群优化算法 特征选择 特征加权 支持向量机 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]