摘要
本文介绍了矿产预测的一种方法——统计决策分析。文中给出了矿产预测条件下统计决策问题的陈述,该问题在先验概率已知和未知时的两种解法:贝叶斯解和极小极大解(对于m>2和m=2两种情况),原始数据预处理的方法。文章通过在三〇七五地区的应用实例展示了该方法的应用效果。
Prediction of minera lresources is a statistical decision problem.Classes of mineralization prospect are regarded as conditions existed in the nature and the predictions of mineralization prospect made by a geologist may be considered as decisions of a decider.In this case the statistical decision problem is stated as follows:given m the classes of mineralization prospect and m,decision of geologists,set X of all possible n-dimensional geological vectors and class{Pi}of their probability distributions for m prospect classes,known or unknown prior probability q of m prospect classes,to find an optional decision functionφcorresponding to some principle,which can guide geologist to select optional decision.In correspondence with two optimality principles are given the Bayes’and minimax solution of statistical deoision problem for the cases of m>2 and m=2 respectively.In the paper the technigue of the first order approximation to class{Pi}of probsbility distributions and methods of pre-proccesing for original data are described.Application of the statistical decision method in the prediction of uranium mineral resources for the 3075 region,for example,is very successful.
作者
殷棣
贺国信
Yin Di;He Guoxin
出处
《地质学报》
1985年第4期319-332,共14页
Acta Geologica Sinica
