Abstract:
The main work of this paper is to study and to achieve a time complexity of low and high precision of clustering data stream clustering algorithm. First is to analyze the data stream mining theory; to analyze and summarize several typical advantages and disadvantages of the traditional clustering algorithms, as well as the scope of application of clustering, which leads to data stream clustering algorithm and it’s elaborate; highlight a new data stream clustering algorithm: the data stream clustering algorithm is based on active mesh density. Firstly, the data space grid is divided into a grid structure formed by the small cube grid cell on the grounds, and then the data stream is mapped to this structure, the application of the concept is of density formation of the concept, and then feature vector to determine the density of the grid. The density attenuation of the dynamic is in the nature of the technology to capture data stream, and then extract the boundary point to remove it; introduce the concept of activity to determine the mesh density of active and inactive grid density to ignore the reserve of active grid density clustering, and in this article the algorithm CluStream algorithm for comparison. Finally, the algorithm is applied in this article to the network intrusion detection algorithm in the detection rate and false alarm rate analyzed to verify whether the algorithm is feasible or not.
Page(s):
209-214
DOI:
DOI not available
Published:
Journal: Journal of Theoretical and Applied Information Technology, Volume: 44, Issue: 2, Year: 2012