ISSN:
1573-8868
Keywords:
nonlinear regression
;
discriminant analysis
;
decomposition of mixtures
;
geochemistry
;
resource assessment
Source:
Springer Online Journal Archives 1860-2000
Topics:
Geosciences
,
Mathematics
Notes:
Abstract It seems unreasonable to use one population to fit the distribution of an element, and then to determine a threshold to separate anomalous data from background data in an analysis of geochemical data. Statistically, anomaly, background, and other geological categories may be represented by different component populations overlapping one another. Therefore, anomaly, background, and other geological categories should be distinguished from one another by distributions rather than by thresholds. This paper uses a method of decomposition of mixtures to identify observed distributions of five elements, obtained in a geochemical reconnaissance of the Silver City-South Mountain region, Idaho, into component populations. Observations have been assigned to populations and mapped; finally, these populations have been interpreted leading to recognition of both mineralized belts and lithologic patterns.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF01034752
Permalink