由粗到精分层技术下的复杂网络入侵检测方法研究

摘要：复杂网络具有开放性、互联性和共享性,易受到大规模的入侵,采用传统“一对一”方式构建网络入侵检测器,检测费时,实时性检测差。为了提高复杂网络入侵检测性能,提出一种引入由粗到精分层概念的多层网络入侵检测模型,在传统的LSSVM分类器基础上,对分类过程进一步细分,建立一种由粗到精策略,构造多层的网络入侵分类器,在精细分类层,将引入拥挤度和隔离度因子的粒子群优化分类器,以提高入侵分类器性能。最后采用KDD 99数据集进行仿真测试。结果表明,相对于其它检测模型,该模型不仅加快了入侵检测速度,满足入侵检测实时性,同时提高了网络入侵检测率,为网络安全提供了有效保证。

Research on coarse-to-fine hierarchical intrusion detection technology in complex network

Abstract: Complex network is open, interconnection and sharing, vulnerable to large-scale invasion, the traditional "one to one" method is used to build network intrusion detection model, which intrusion detection is time consuming, effects of real-time for intrusion detection is poor, in order to improve the performance of network intrusion detection, this paper proposed network intrusion model based on a coarse-to-fine hierarchical technical. Based on traditional LSSVM classifier, the classification process is further subdivided, a coarse-to-fine strategy is used to establish a network multilayer structure intrusion classifier, in fine classification layer, the crowded and isolation factor particle swarm classifier is introduced to improve the performance of the classifier,. Finally, simulation test is carried out by using KDD 99 data sets. The results show that, compared with other detection model, this proposed model can not only accelerate the speed of intrusion detection, can meet the real-time requirement of network intrusion detection, and improve the network intrusion detection rate, and it can provide effective guarantee for network security.