Research On Intrusion Detection Technology Based On Improved K-means Algorithm

Nowadays,network applications have gained tremendous popularity,and the resulting network security issues have become the focus of attention.In order to solve the problems caused by network security,researchers have proposed many methods and technologies.Among them,intrusion detection technology has gradually become a research hotspot,and many research results have been obtained.However,the traditional intrusion detection technology is difficult to meet the current complex and changeable Internet environment.There are shortcomings such as false negatives and high false positive rates,which cannot meet the actual needs.Therefore,it is a difficult and hot issue to find an efficient and accurate intrusion detection technology.Based on the research of the principles,basic processes and defects of intrusion detection,this paper improves K-means algorithm and designs a new detection method.Experimental results show that the improved detection algorithm has better results and effectively solves the shortcomings of traditional algorithms.The main research of this article is as follows:(1)Research the classic algorithm K-means in cluster analysis in detail,analyze its basic principle and clustering process,and summarize some remaining shortcomings of K-means algorithm: weak anti-noise ability,sensitive initial centroid,many iterations,and human Specify the number of clusters,etc.(2)In view of the above deficiencies,the K-means algorithm is optimized and an improved algorithm is proposed.In the improved algorithm,the Canopy algorithm is introduced,and a coarse clustering is performed before the K-means algorithm,and then uses "anti-noise","maximum minimum principle","density maximum principle",depth theory and K-means calculation simplification strategy to optimize the algorithm.(3)Experimental research on the effectiveness of the improved algorithm.Randomly generated two-dimensional data sets are clustered using the algorithms before and after improvement,and their SSE values are calculated separately.The experimental results show that compared with the traditional K-means algorithm,the improved K-means algorithm obviously shows better clustering effect.(4)The KDD CUP99 data set is used to conduct intrusion detection simulation experiments.The experimental data demonstrate that the false positive rate is reduced and the detection rate of intrusion detection is increased by the improved clustering algorithm,which effectively improves the comprehensive performance of intrusion detection.