Guidelines for Enhancing the Signature of Multi-element Mineralization Using Principal Component Analysis: Part 1—Monte Carlo Simulation
- 41 Downloads
Principal component analysis (PCA) is a widely used method in geochemical data processing. The method can be useful to integrate variables associated with mineralization into a single component. In this paper, a Monte Carlo simulation is designed and applied to explore the performance of PCA under conditions controlled by four factors: the number of geo-objects (lithologic units), differences between geo-objects, the relationship between the variables and the number of variables. The results imply that: (1) more significant differences between geo-objects will result in less stable PC results; (2) more geo-objects make the result more robust; (3) variables with similar relationships help to stabilize the result; (4) more input variables do not always lead to a better result. These conclusions provide useful guidelines for using PCA to yield a targeted component like mineralization.
KeywordsPrincipal component analysis Monte Carlo simulation Data mining Geochemical exploration
We are thankful for the suggestions and the modifications from an anonymous reviewer and from Prof. Graeme Bonham-Carter. This research benefited from financial support from National Key R&D Program of China (2016YFC0600501), National Natural Science Foundation of China (Nos. 41430320 and 41602337) and a Chinese Geological Survey project (Minerals and Geological Prospecting on Shallow Covered Areas of Jinning, Inner Mongolia, No. DD20160045).
- Ball, T. K., Brown, M. J., Nicholson, R. A., Peachey, D., & Smith, T. K. (1984). Comparison of different geochemical prospecting techniques over the Long Rake, Fluorite–Barite–Sulfide Orebody, Derbyshire. Journal of the Geological Society, 141, 390–390.Google Scholar
- Carranza, E. J. M. (2008). Geochemical anomaly and mineral prospectivity mapping in GIS (1st ed.). Amsterdam: Elsevier.Google Scholar
- Cheng, Q., Bonham-Carter, G., Wang, W., Zhang, S., Li, W., & Qinglin, X. (2011). A spatially weighted principal component analysis for multi-element geochemical data for mapping locations of felsic intrusions in the Gejiu mineral district of Yunnan, China. Computers & Geosciences, 37, 662–669.CrossRefGoogle Scholar
- Grunsky, E. C., & Kjarsgaard, B. A. (2016). Recognizing and validating structural processes in geochemical data: Examples from a diamondiferous kimberlite and a regional lake sediment geochemical survey. In J. A. Martín-Fernández & S. Thió-Henestrosa (Eds.), Compositional data analysis: CoDaWork, L’Escala, Spain, June 2015 (pp. 85–115). Cham: Springer International Publishing.CrossRefGoogle Scholar
- Harris, D., & Pan, G. (1999). Mineral favorability mapping: A comparison of artificial neural networks, logistic regression, and discriminant analysis. Nonrenewable Resources, 8, 93–109.Google Scholar
- Liu, B., Guo, K., & Zhang, L. (2016). Kernel principal component analysis in the application of geochemical comprehensive feature extraction (pp. 15–19). Cham: Springer International Publishing.Google Scholar
- Pirajno, F. (2012). Hydrothermal mineral deposits: Principles and fundamental concepts for the exploration geologist. Berlin: Springer.Google Scholar
- Roweis, S. (1998). EM algorithms for PCA and SPCA. In M. I. Jordan, M. J. Kearns, & S. A. Solla (Eds.), Advances in neural information processing systems (Vol. 10, pp. 626–632). Cambridge: MIT Press.Google Scholar
- Sawilowsky, S. S. (2003). You think you’ve got trivials? Journal of Modern Applied Statistical Methods, 2, 21.Google Scholar
- Thomopoulos, N. T. (2012). Essentials of Monte Carlo simulation: Statistical methods for building simulation models. New York: Springer.Google Scholar