摘要
针对物联网应用实时性强和资源受限的特点,提出了一种基于改进聚类分析和智能优化特征提取的物联网入侵检测算法。首先,构建基于模糊聚类和特征提取的入侵检测模型,训练阶段对样本数据进行聚类分析,并利用特征提取算法获得每个分类最佳特征子集;入侵检测阶段在判定测试数据最佳聚类归属的基础上,利用该分类对应的特征子集与测试数据进行对比检测,从而实现了物联网入侵高效率和高准确率检测。其次,设计最佳聚类指数并引入模糊C均值聚类算法(FCM)中,以实现FCM聚类个数自动划分;设计基于离散混合蛙跳算法的特征提取算法,以实现最佳特征子集提取。最后,相比于其检测算法,入侵检测效率和准确率较为改善。
Aiming at the real-time and resource-constrained characteristics of the Internet of Things( IOT ), an intrusion detection algorithm based on improved clustering analysis and intelligentoptimization feature extraction is proposed. Firstly, an intrusion detection model based on fuzzy clusteringand feature extraction is constructed. In training phase, the sample data is clustered and analyzed, andthe best feature subset of each classification is obtained by feature extraction algorithm. The test data arecompared and detected to achieve high efficiency and high accuracy detection of Internet intrusion.Secondly, the optimal clustering index is designed and the fuzzy C-means clustering algorithm (FCM) isintroduced to realize the automatic division of the number of FCM clusters, and the feature extractionalgorithm based on discrete shuffled frog leaping optimization is designed to achieve the best featuresubset extraction. Finally, compared with others detection algorithms, the efficiency and accuracy ofintrusion detection are improved.
作者
张麾军
ZHANG Hui-jun(China Mobile Communications Group Chongqing Co.,Ltd.,Chongqing 401122,China)
出处
《信息技术》
2018年第12期103-107,共5页
Information Technology
关键词
物联网
入侵检测
模糊聚类
特征提取
离散混合蛙跳算法
Internet of Things
intrusion detection
fuzzy clustering
feature extraction
discreteshuffled frog leaping optimization