论文部分内容阅读
K型识别表法是处理定性地质资料而设计的进行二类判别的方法。文中简要介绍了K型识别表的概念,标志分类权的计算方法,对象间海明距离的定义和对未知对象进行判别的准则。为了在该方法中也能利用连续变量的资料,介绍了连续变量离散化的信息度最大的方法。最后,用作者对该方法所设计的源程序对我国南方二十一个未知归属的花岗岩体进行计算试验,确定了六个具有远景的岩体。
K-type identification table method is to deal with qualitative geological data designed to carry out two types of discrimination method. In this paper, the concept of K type identification table, the method of calculating the classification weight of sign, the definition of Heming distance between objects and the criterion of discriminating unknown objects are briefly introduced. In order to make use of the data of continuous variables in this method, the method of maximizing the degree of information discretization of continuous variables is introduced. Finally, with the source code designed by this method, twenty-six unknown granite bodies in southern China were calculated and tested, and six rock masses with long-term perspectives were determined.