GEOSCAN Search Results: Fastlink


TitleRobust principal component analysis for power transformed compositional data
AuthorScealy, J L; de Caritat, P; Grunsky, E C; Tsagris, M T; Welsh, A H
SourceJournal of the American Statistical Association vol. 110, issue 509, 2015 p. 136-148, Open Access logo Open Access
Alt SeriesEarth Sciences Sector, Contribution Series 20140228
PublisherInforma UK Limited
Mediapaper; on-line; digital
File formatpdf
ProgramGEM: Geo-mapping for Energy and Minerals Geomapping for Energy & Minerals (GEM) - Program Coordination
Released2015 01 07
AbstractGeochemical surveys collect sediment or rock samples, measure the concentration of chemical elements and report these typically either in weight percent orin parts per million. There are usually a large number of elements measuredand the distributions are often skewed, containing many potential outliers. Wepresent a new robust principal component analysis (PCA) method for geochemical survey data, that involves ?rst transforming the compositional data ontoa manifold using a relative power transformation. A ?exible set of moment assumptions are made which take the special geometry of the manifold into account. The Kent distribution moment structure arises as a special case when the chosen manifold is the hypersphere. We derive simple moment and robust estimators of the parameters which are also applicable in high dimensional settings. The resulting PCA based on these estimators is done in the tangent space and is related to the power transformation method used in correspondence analysis. To illustrate, we analyse major oxide data from the National Geochemical Survey of Australia. When compared with the traditional approach in the literature based on the centred logratio transformation, the new PCA method is shown to be more successful at dimension reduction and gives interpretable results.
Summary(Plain Language Summary, not published)
The paper describes the use of advanced statistics for evaluating multi-element geochemical data for more effective pattern interpretation. These patterns are associated with variability in bedrock composition and the potential to identify sites with potential precious and/or base metal mineralization.

Date modified: