基于MST的分布式分簇算法

Distributed hierarchical clustering algorithm based on MST

为了使分簇后的网络更便于数据融合,对最小生成树(MST)的性质进行了研究,论证并实现了一种新的基于MST性质的分布式多层分簇算法。分簇过程中,节点各自独立运行该算法,利用生成的局部MST传递并融合连接信息以完成本层级的网络分簇。经过多次的连接信息间的融合,逐渐形成一个便于数据融合的多层分簇网络。实验分析表明,该算法具有收敛速度快、资源消耗低的优点。

To make clustering network more convenient to aggregate data, the research on the nature of the minimum spanning tree （MST） is made and a new distributed hierarchical clustering algorithm which is based on the nature is demonstrated and designed. During the course, each of the nodes runs the algorithm independently and in order to complete sub-level clustering nodes transmit and aggregate link information via the generated local MST. With several times of aggregations of link information, a hierarchical clustering network which facilitates data fusion comes into being gradually. Experimental analysis shows that this algorithm has the advantages of high convergence speed and low consumption of resources.