Abstract
Most regional geochemistry data reflect processes that can produce superfluous bits of noise and, perhaps, information about the mineralization process of interest. There are two end-member approaches to finding patterns in geochemical data—unsupervised learning and supervised learning. In unsupervised learning, data are processed and the geochemist is given the task of interpreting and identifying possible sources of any patterns. In supervised learning, data from known subgroups such as rock type, mineralized and nonmineralized, and types of mineralization are used to train the system which then is given unknown samples to classify into these subgroups.
To locate patterns of interest, it is helpful to transform the data and to remove unwanted masking patterns. With trace elements use of a logarithmic transformation is recommended. In many situations, missing censored data can be estimated using multiple regression of other uncensored variables on the variable with censored values.
In unsupervised learning, transformed values can be standardized, or normalized, to a Z-score by subtracting the subset's mean and dividing by its standard deviation. Subsets include any source of differences that might be related to processes unrelated to the target sought such as different laboratories, regional alteration, analytical procedures, or rock types. Normalization removes effects of different means and measurement scales as well as facilitates comparison of spatial patterns of elements. These adjustments remove effects of different subgroups and hopefully leave on the map the simple and uncluttered pattern(s) related to the mineralization only.
Supervised learning methods, such as discriminant analysis and neural networks, offer the promise of consistent and, in certain situations, unbiased estimates of where mineralization might exist. These methods critically rely on being trained with data that encompasses all populations fairly and that can possibly fall into only the identified populations.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
REFERENCES
Bonham-Carter, G. F., 1994, Geographic information systems for geoscientists: Modelling with GIS: Pergamon, New York, 398 p.
Brown, W. M., Gedeon, T. D., Groves, D. I., and Barnes, R. G., 2000, Artificial neural networks: A new method for mineral prospectivity mapping: Australian Jour. Earth Sciences, v. 47, no. 4, p. 757–770.
Cheng, Q., 1999, Spatial and scaling modeling for geochemical anomaly separation: Jour. Geochem. Exploration, v. 65, no. 3, p. 175–194.
Cheng, Q., Xu, Yaguang, and Grunsky, E., 2000, Integrating spatial and spectrum method for geochemical anomaly separation: Natural Resources Research, v. 9, no. 1, p. 43–51.
Doyle, A. C., 1888, The sign of four: Hardcover reprint edition (December 1993), Applewood Books, Bedford, Massachusetts, 112 p.
Eggo, A. J., 1996, Regional geochemistry: a continental perspective (abst.), in Kennard, J. M., ed., 13th Australian Geol. Conv., AGSO Jubilee Symposium, Canberra: Geol. Soc. Australia Abstracts, v. 41. p. 126.
Gendall, I. R., Quevedo, L. A., Sillitoe, R. H., Spencer, R. M., Puente, C. O., Leόn, J. P., and Povedo, R. R., 2000, Discovery of a Jurassic porphyry copper belt, Pangui area, southern Ecuador: SEG Newsletter (Oct, 2000), no. 43, p. 1–15.
Grunsky, E. C., 1986, Recognition of alteration in volcanic rocks using statistical analysis of lithogeochemical data: Jour. Geochem. Exploration, v. 25, no. 1, p. 157–183.
Gustavsson, N., and Kontio, M., 1990, Statistical classification of regional geochemical samples using local characteristic models and data of the geochemical atlas of Finland and from the Nordkalott Project, in Gaál, G., and Merriam, D. F., eds., Computer Applications in Resource Estimation—Prediction and Assessment for Metals and Petroleum: Pergamon Press, Oxford, p. 23–41.
Harris, J. R., Wilkinson, L., Grunsky, E., Heather, K., and Ayer, J., 1999, Techniques for analysis and visualization of lithogeochemical data with applications to the Swayze greenstone belt, Ontario: Jour. Geochem. Exploration, v. 67, nos. 1–3, p. 301–334.
Howarth, R. J., ed., 1983, Statistics and data analysis in geochemical prospecting: Handbook of exploration geochemistry, vol. 2, Elsevier, New York, 437 p.
Kotlyar, B. B., Singer, D. A., Jachens, R. C., and Theodore, T. G., 1998, Regional analysis of the distribution of gold deposits in northeast Nevada using NURE arsenic data and geophysical data, in Tosdal, R. M., ed., Contributions to the Au Metallogeny of the Northern Great Basin: U.S. Geol. Survey Open-File Rept. 98–338, p. 234–242.
Marcotte, D., and David, M., 1981, Rtarget-definition of Kurokotype deposits in Abitibi by discriminant analysis of geochemical data: CIM Bull. v. 74, no. 828, p. 102–108.
Masters, T., 1993, Practical neural network recipes in C++: Academic Press, San Diego, California, 493 p.
Rehder, S., and Van Den Boom, G., 1982, Geochemical characterization of tin granites in Northern Thailand, in Howarth, R. J., ed., 1983, Statistics and Data Analysis in Geochemical Prospecting: Handbook of exploration geochemistry, vol. 2, Elsevier, New York, p. 311–339.
Sanford, R. F., Pierson, C.T., and Crovelli, R. A., 1993,Anobjective method for censored geochemical data: Math. Geology, v. 25, no. 1, p. 59–80.
Sinclair, A. J., 1974, Selection of thresholds in geochemical data using probability graphs: Jour. Geochem. Exploration, v. 3, no. 1, p. 129–149.
Singer, D. A., and Kouda, R., 1988, Integrating spatial and frequency information in the search for kuroko deposits of the Hokuroku District, Japan: Econ. Geology, v. 83, no. 1, p. 18–29.
Singer, D. A., and Kouda, R., 1991, Application of the FINDER system to the search for epithermal vein gold-silver deposits: Kushikino, Japan, a case study: Geoinfomatics, v. 2, no. 2, p. 113–123.
Singer, D. A., and Kouda, R., 1996, Application of a feedforward neural network in the search for kuroko deposits in the Hokuroku District, Japan: Math. Geology, v. 28, no. 8, p. 1017–1023.
Singer, D. A., and Kouda, R. 1997a, Classification of mineral deposits into types using mineralogy with a probabilistic neural network: Nonrenewable Resources, v. 6, no. 1, p. 27–32.
Singer, D. A., and Kouda, R., 1997b, Use of a neural network to integrate geoscience information in the classification of mineral deposits and occurrences, in Gubins, A. G., ed., Proc. Exploration 97: Fourth Decennial Intern. Conf. Mineral Exploration (Toronto, Canada), p. 127–134.
Singer, D. A., and Kouda, R., 1999, A comparison of the weights of evidence method and probabilistic neural networks: Natural Resources Research, v. 8, no. 4, p. 287–298.
Theodore, T. G., Kotlyar, B. B., Berger, V. I., Moring, B. C., Singer, D. A., and Edstrom, S. A., 1999, Geochemistry of streamsediment samples from the Santa Renia Fields and Beaver Peak quadrangles, northern Carlin trend, Nevada: U.S. Geol.Survey Open-File Rept. 99–341, 103 p.
Theodore, T. G., Kotlyar, B. B., Berger, V. I., Moring, B. C., and Singer, D. A., 2000a, Implications of stream-sediment geochemistry in the northern Carlin trend, Nevada, in Cluer, J. K., Price, J. G., Struhsacker, E. M., and Morris, C. L., eds., Geology and Ore Deposits 2000:The Great Basin and Beyond: Geol. Soc. Nevada Symp. Proc. (May 15–18, 2000), p. 929–958.
Theodore, T. G., Kotlyar, B. B., Moring, B. C., Singer, D. A., and Edstrom, S. A., 2000b, Geochemistry of rock samples from the Santa Renia fields and Beaver Peak quadrangles, northern Carlin trend, Nevada: U.S. Geol. Survey Open–File Rept. 00–402, 125 p.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Singer, D.A., Kouda, R. Some Simple Guides to Finding Useful Information in Exploration Geochemical Data. Natural Resources Research 10, 137–147 (2001). https://doi.org/10.1023/A:1011552810482
Issue Date:
DOI: https://doi.org/10.1023/A:1011552810482