The multivariate methods, useful for prospect appraisal and calibration are the following:

- Multiple regression

This method refers to a regression analysis where a dependent data (e.g. reserves) is related to a set of independent (e.g. geological) variables, in order to predict reserves from geological factors. - Multiple Regression with some
**censored values**for the dependent variable.

Same as the multivariate method, but allows data of the depndent variable to be truncated. This means that some of the data are only minimum values. The true value may be larger but cannot be observed, a common feature in geological data. - Discriminant analysis.

A statistical technique that tries to find a way to find a function of several variables that helps to classify multivariate objects into two or more classes. We could think of "dry holes" versus "discoveries". - Clustering methods by non-linear mapping

A sample of multivariate data can be analysed for clustering. It is possible that data may not form a single "cloud" of points, but falls apart into two or more "clusters" that can be related to certain geological characteristics.