Data Streams Clustering Algorithm Based on Grid and Particle Swarm Optimization
The offline components of CluStream clustering algorithm based on distance, and it is difficult to find non-spherical character of the cluster. This paper proposes a data streams clustering algorithm based on grid and particle swarm optimization, the algorithm based on two-tier structure of CluStream clustering algorithm. The grid feature vector to represent a snapshot, the grid density, and grid merging technologies is applied in this algorithm. So we confound any non-spherical clusters,In this algorithm.Non-dense grid is periodic and dynamic way to remove,which is good for reducing the space complexity.Using the PSO optimized clustering results in the offline components, in order to get a more precise clustering efficiency. Experiments show that this algorithm is more efficient than the CluStream algorithms, it has a good number of dimensions scalability,and it can find non-spherical nature of the clustering results.
Grid density Data Stream Particle Swarm Optimization Clustering Algorithm
Luo Ke Wang Lin
School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha, 410076,China
国际会议
重庆
英文
93-96
2009-12-25（万方平台首次上网日期，不代表论文的发表时间）