Some statistical properties of the minimum spanning forest
作者:
Highlights:
•
摘要
A minimum spanning forest method is described which finds cluster patterns in a random graph of points. A relation between the main number of clusters and a defined fixed thresholding value is obtained. Then, using the number of clusters as a parameter, a uniformity test suitable for low statistics data sets is proposed. Finally, some effective applications on simulated and real data are also discussed.
论文关键词:Clustering,Image analysis,Preprocessing,Discriminant analysis,Uniformity test
论文评审过程:Received 6 July 1981, Revised 17 March 1983, Accepted 8 April 1983, Available online 19 May 2003.
论文官网地址:https://doi.org/10.1016/0031-3203(83)90057-2