基于机器学习算法的网络入侵检测  

Network intrusion detection based on machine learning algorithm

作　　者：张夏 

机构地区：[1]宜春学院,江西宜春336000

出　　处：《现代电子技术》2018年第3期124-127,共4页Modern Electronics Technique

摘　　要：网络入侵的频率越来越高,严重危害了网络安全。为了获得高正确率的网络入侵检测结果,针对当前网络入侵检测模型的局限性,提出基于机器学习算法的网络入侵检测模型,通过机器学习算法中性能优异的支持向量机构建"一对一"的网络入侵检测分类器,采用当前标准网络入侵检测数据库对模型的有效性进行验证,网络入侵检测正确率高达95%以上,检测误差远远低于实际应用范围,可以应用于实际的网络安全管理中。The frequent network intrusion endangers the network security seriously.In order to obtain the network intrusion detection results with high accuracy,a network intrusion detection model based on machine learning algorithm is proposed for the limitations of the current network intrusion detection model.The support vector machine of machine learning algorithm is used to construct the one-to-one network intrusion detection classifier.The standard network intrusion detection database is used to verify the effectiveness of the model with experiment.The network intrusion detection rate is higher than 95%,the detection error is far below the actual application range.The model can be applied to the practical network security management.

关 键 词：网络安全 入侵行为 机器学习算法 入侵检测 分类器 检测误差 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP393.08[自动化与计算机技术—控制科学与工程]