RI 6876 A Computer Method Of Fitting Surfaces To Assay And Other Data In Three Dimensions By Quadratic-Regression Analysis

A statistical technique, useful to analyze assay and other data characterized by location in three dimensions, is described in this Bureau of Mines report. If present, trends in the data are defined for the purposes of improving ore estimation, predicting ore beyond existing workings, and aiding geological interpretation. The technique is to fit linear and quadratic polynomial equations to the data through the statistical method of regression analysis. The fitted linear equations have the geometric form of hyperplanes; the fitted quadratic equations have the geometric form of ellipsoids or elliptic hyperboloids of one or two sheets. To visualize these three-dimensional forms, contour maps of a series of plane sections through the fitted surfaces are plotted. Both the fitting of the polynomial equations and the plotting of the con-tour maps are done by electronic computer to facilitate calculations and to obtain accurate results. Computer programs are described in appendixes. Methods employed are illustrated by an example analysis) based on data from a part of the Fresnillo mine) Zacatecas, Mexico.