On the use of orthogonally stepwise regression discrimination for predicting copper ore deposits of Dongchuan type in central Yunnan, China

In the prediction of ore deposits, a series of geological, geochemical, and geophysical characters are used to describe the model of ore deposits. The problem is how to use the characters of known ore deposits to establish the model and to look for the target areas with similar geological environments. This can be done with a powerful mathematical tool—discriminant analysis. Nevertheless, there generally are too many geological variables to describe a class of ore deposits. Some of them are discrete, and the others are continuous. It is also difficult to know their distributions. A satisfactory method of discrimination has not been developed for such a complicated case. For this reason, we introduce a new method of orthogonally stepwise discrimination. We used this method to predict copper ore deposits of Dongchuan type in central Yunnan, China, with satisfactory results.