WSN中基于Mini Batch K-Means与SVM的入侵检测方案

An Intrusion Detection Scheme Based on Mini Batch K-Means and SVM in Wireless Sensor Networks

无线传感器网络通常部署在复杂的户外环境,易遭受各种攻击。多数入侵检测系统均采用数据挖掘算法对网络数据包进行分析,但在处理大样本集时,其效率明显降低。针对这一缺点,提出一种基于Mini Batch K-Means和SVM的入侵检测方案。该方案首先分别对正常行为特征库和异常行为特征库进行Mini Batch K-Means聚类,取得类中心作为各类的代表样本并赋予权值,将其传入SVM分类器作为训练数据,得到分类超平面,通过该超平面对待测样本作出判断。解决了如K-Means、KNN、SVM等传统数据挖掘算法在大数据样本集数据分析中面临的低效问题。仿真结果表明,该方案能快速准确地判断样本类别,其检测率达到98.7%。与K-Means、KNN和SVM相比,不仅达到了同样高的检测率,而且明显提高了入侵检测的时间效率。

Wireless sensor networks(WSN)are usually deployed in complex outdoor environments and vulnerable to various attacks.Most intrusion detection systems use data mining algorithms to analyze network packets,but their efficiency is significantly reduced when dealing with large sample sets.In view of this shortcoming,a intrusion detection scheme based on Mini Batch K-Means and SVM is proposed.The scheme firstly conducts Mini Batch K-Means clustering on the database of normal behavior and abnormal behavior.The obtained centers of clusters are then used as representative samples of their clusters,each of which is assigned a weight and is input into SVM classifier as training data,so as to get a Classifying hyper-plane which can be used to classify samples.The proposed scheme solves the inefficiencies that traditional data mining algorithms such as K-means,KNN and SVM face in data analysis of big data sample sets.Simulation results show that our proposed scheme can determine the sample types quickly and accurately,whose accuracy is 98.7%.Compared with K-Means,KNN and SVM,our proposed scheme not only achieved the same accuracy,but also improved the time efficiency of intrusion detection significantly.