Real-time data stream clustering and its boundary detection based on distance and density
Xiaolong Zhang, Xiaobo Liang, Bo Li · 2011
The real-time data stream clustering and the detection of clustering boundary is an interesting research work. This paper proposes a clustering algorithm named DDBound with boundary detection ability for grid clustering based on distance and density. DDBound firstly calculates the densities of all grids, and divides them into high-density grids and low-density grids. From all grids, the algorithm repeatedly finds the maximal density grid which has not been clustered. Begin with this grid, the depth-first traversal is used to make the high-density grids be connected with each other, and transitional grids connected with the high-density grids into a cluster. Finally, boundary grids are extracted from the clustering results. The experimental results demonstrate that DDBound algorithm can effectively identify clusters in data stream of arbitrary shapes, sizes and different densities and find the boundary of clusters.