摘要
自组织特征映射是一种无监督的神经网络,目前广泛应用于入侵检测中。文章提出了一种基于改进的SOM的入侵检测方法,可更有效的处理包含数值型和字符型的输入向量,优化了训练中的权值调整策略。最后,使用KDD Cup 99数据集进行实验,结果表明改进的SOM算法检测率较高。
As an unsupervised neutral network,Self-Qrganizing Map(SOM) has been used widely in intrusion detection.This paper proposes an improved SOM for detecting intrusions,which can deal with numerical and symbolic data more effectively,with weight adjustment optimized.The KDD Cup 99 dataset is used in ualidity experiments,whose results indicate that the improved SOM leads to a higher detection rate.
出处
《信息工程大学学报》
2011年第5期630-633,共4页
Journal of Information Engineering University
基金
河南省重点科技攻关项目(082102210097)
关键词
自组织特征映射
入侵检测系统
聚类
self-organizing map
intrusion detection system
clustering