Intrusion Detection Model with Twin Support Vector Machines

来源 :Journal of Shanghai Jiaotong University(Science) | 被引量 : 0次 | 上传用户:leaf678
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
Intrusion detection system(IDS) is becoming a critical component of network security. However,the performance of many proposed intelligent intrusion detection models is still not competent to be applied to real network security. This paper aims to explore a novel and effective approach to significantly improve the performance of IDS. An intrusion detection model with twin support vector machines(TWSVMs) is proposed.In this model, an efficient algorithm is also proposed to determine the parameter of TWSVMs. The performance of the proposed intrusion detection model is evaluated with KDD’99 dataset and is compared with those of some recent intrusion detection models. The results demonstrate that the proposed intrusion detection model achieves remarkable improvement in intrusion detection rate and more balanced performance on each type of attacks.Moreover, TWSVMs consume much less training time than standard support vector machines(SVMs). However, the performance of many proposed intelligent intrusion detection models is still not competent to be applied to real network security. This paper aims to explore a novel and effective approach to significantly improve the performance of IDS. An intrusion detection model with twin support vector machines (TWSVMs) is proposed.In this model, an efficient algorithm is also proposed to determine the parameter of TWSVMs. The performance of the proposed intrusion detection model is evaluated with KDD ’99 dataset and is compared with those of some recent intrusion detection models. The results demonstrate that the proposed intrusion detection model achievious remarkable improvement in intrusion detection rate and more balanced performance on each type of attacks.Moreover, TWSVMs consume much less training time than standard support vector machines (SVMs).
其他文献
植被物候指植物在气候和环境因子影响下而发生的周期性自然现象,其变化是气候与环境变化最敏感且易观测的综合标志。植被净初级生产力(NetPrimaryProductivity,NPP)是植物单位
随着课改的深入,一部分敢于、善于钻研和思考的教师把课改视角从课堂教学转向了课程设置。对于课程设置的研究,我们比较熟悉的是三级课程设置,即国家课程、地方课程及校本课
期刊
对单片机应用电路进行简要分析,介绍双音频拨号的编程思路,重点分析用户卡读写电路的基本原理和程序工作过程。 Briefly analyze the application circuit of SCM, introduc
通过合成醋酸乙烯酯(VAc)、丙烯酸丁酯(nBA)、甲基丙烯酸甲酯(MMA)、丙烯酸(AA)四元共聚合乳液,并添加陶瓷粉,制成高分子保护膜,将该膜均匀涂于寒富苹果1年生枝条及顶花芽进
随着全球城市化进程的加快,城市热岛效应愈加显著,如何定量监测、分析与评价城市热岛效应,已成为当前城市气候与环境研究中,最为重要的研究内容之一。因此,深入研究城市热岛在空间
地表温度(Land Surface Temperature,LST)是表征地球表面热力学状态的物理参数,是地-气之间长波辐射和湍流热通量的直接驱动因素,是区域与全球尺度地表能量和水循环过程的关键指
公路运输作为交通运输体系中的重要组成部分,在推动经济社会发展中发挥着基础性作用。改革开放以来,伴随着经济的高速发展,我国公路基础设施建设取得了举世瞩目的成就,形成了较为
不透水面(Impervious Surface Area,ISA)是一种常见的人工建筑地表特征,经常被作为评价区域生态环境健康的重要指标,也被广泛的应用在区域城市扩张监测研究中。不透水面不仅对区
2016年,中国主要淡水湖中太湖、巢湖、洪泽湖和滇池均处于轻度乃至中度富营养化。蓝藻的暴发性繁殖,是湖泊水体富营养化的重要特征,中国已经成为世界上蓝藻水华暴发最严重、分布
该文围绕遥感数据专题分类的不确定性问题和遥感信息的尺度问题,具体做了以下几个方面的工作:(1)针对传统的分类不确定性评价法不能反映不确定性空间变化的缺陷,以概率矢量为基