BP神经网络在入侵检测中的应用  

The Application of BP-Neural Network in the Intrusion Detection

作　　者：包晖[1] 秦杰[1] 张兴红[1] 

机构地区：[1]河南工业大学信息科学与工程学院,河南郑州450052

出　　处：《计算机安全》2010年第3期22-24,共3页Network & Computer Security

基　　金：国家自然科学基金项目;编号:60373003;郑州市科技攻关项目;编号:083SGYG25121-7

摘　　要：针对传统入侵检测系统在误报率、漏报率上的缺陷,把神经网络应用于专业入侵检测系统中,利用神经网络的BP算法的学习能力和快速识别能力,实现了对用户行为的检测,最后给出了该算法的理论推导和算法的具体实现步骤。与传统的入侵检测算法相比,该方法的最小平均误差比传统算法的最小平均误差小,且学习速度快。实验结果也说明了该方法的快速性、有效性、稳定性。In allusion to the insufficiencies such as missing report rate and false alarm rate, the paper applys neural network into intrusion detection, achieves a goal of user behavior with BP algorithm is learning ability and fast recognition ability. At last the paper gives out a theoretical derivation and the steps of algorithm. Compared with old algorithm, new algorithm＇s minimum average error is smaller than the old algorithm＇s, furthermore, its learning speed is faster. From the results of experiment, we can see that the new algorithm has some advantages, such as quickly, validity and practicability.

关 键 词：BP 神经网络 IDS 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]