本文着重介绍地质多元统计分析在矿产预测中的应用.该方法始于六十年代,但大规模地开展是从七十年代开始,直接用于矿床预测的多元统计分析方法有回归分析、判别分析、聚类分析、因子分析及对应分析等.这些方法的应用在国内外都取得了不同程度的效果.现重点介绍前三种多元统计分析方法在矿床预测方面的应用. This article focuses on the application of geo-multivariate statistical analysis in mineral prediction, which began in the 1960s, but on a large scale since the 1970s, the multivariate statistical methods that are directly used for ore deposit prediction include regression analysis and discriminant analysis , Cluster analysis, factor analysis and correspondence analysis, etc. The application of these methods has achieved some effects both at home and abroad.The applications of the first three kinds of multivariate statistical analysis methods in ore deposit prediction are highlighted.