基于优化网格的最小生成树聚类算法

提出一种基于优化网格的最小生成树聚类算法OGMST(MST Clustering Algorithm Based on Optimized Grid),一方面利用最小生成树的方法进行聚类,另一方面利用了参数自动化的网格划分技术和密度阈值处理技术,解决了最小生成树聚类算法不适用于多密度数据集的局限性,提高了现有的多密度聚类算法的效率和精度.算法还对边界点进行了有效的处理.实验表明算法具有较好的可扩展性,是一种高效、快速的聚类算法.

MST CLUSTERING ALGORITHM BASED ON OPTIMIZED GRID

An MST clustering algorithm based on optimized gird(OGMST) was presented.While carrying on clustering in the way of MST,the OGMST resolved the MST algorithm's limitation in unfitting for multi-density datasets by the use of parameter automatic grid partition technique and density threshold method,and improved the efficiency and precision of the existent clustering algorithms on multi-density datasets.Besides,the OGMST can extract border points effectively.The experiment results show that the OGMST is of goo...