Research Of Improving For K-means Clustering Algorithm

In the United States military to develop the Internet,and open to the outside world,the rapid development of computer systems,more and more scientists,researchers put into the development of computer systems.At the same time,the electronic data has become the mainstream in the Internet and independent computer systems to produce more and more data,more and more redundant data,how to find the massive data we need to become With the development of the Internet an important research topic,the nature of the study is the mining of data summarized,and clustering algorithm is developed in the development process is a very important method,clustering algorithm research history is very long,each step Development has its own unique characteristics,in the identification of data has important significance.The traditional K-means clustering algorithm is not perfect in applicability.With the complexity and dimensionality of the data object,the traditional K-means clustering algorithm is not perfect.In this paper,,The complexity of the problem appears to rise.Based on the traditional k-means algorithm,an improved clustering algorithm is proposed to solve the problem of traditional clustering algorithm.The intrusion detection technology is applied to the traditional algorithm and the improvement of clustering algorithm.The results show that the improved algorithm is effective.