对似近分析的一点改进及其拓广

AN IMPROVEMENT ON RESEMBLANCE-NEAR ANALYSIS AND ITS GENERALIZATION

本文给出了似近分析一个重要的改进，并将改进的似近分析推广为m维似近分析，同时给出了两个数学定理及其证明。根据多维似近系数的数学性质，提出了一种新的聚类方法，即逐步多维聚类法。实例分析证明，这种聚类法比二维系数聚类优越，聚类功能强，所获得的结果更为客观。

An improyement on resemblance=near analysis was made and then the improved resemblance-near analysis was generalized to become a resemblance-near analysis of m dimensions,at the same time,two theorems in the resemblance-near analysis were drawn from this improvement,and proofel in mathematics in this article.On the basis of the nature of resemblance-near coefficient of multiple dimensions,a new cluster method---the cluster used by parameters,(A) of step b/step muliple dimensions was put foreward.Example analysis showed that this cluster method has an advance over the one used by parameter,l(2),of two dimensions in cluster function and objective degree of cluster tree.