基于局部单元特性提取的聚类算法研究

Research on Clustering Algorithm Based on Local Cells Feature Extraction

该算法思想将整体区域划分成若干局部单元区域,并将这些单元标以密集与否、相互间存在阻碍与否等统计特性,然后,通过分析局部单元特性找出相互间有关联的、密集的、不存在阻碍的单元,基于该相对密集区域设立新的聚类中心点并返回。最后,通过分析该算法复杂度得出该算法具有较少输入参数、结果的可靠性等优越性。

The proposed algorithm divides the overall area into a number of local cells. Each cell is associated with statistical information that enables us to label the cell as dense or non-dense , also label each cell as obstructed or non-obstructed. Then the algorithm finds the regions of connected, dense, non-obstructed cells. the algorithm founds a new cluster center for each such region and returns those centers as centers of the relatively dense regions. Finally, the algorithm requires less input parameters and the quality of results is guaranteed according to its complexity analysis.