Search

R-mode hierarchical clustering is a method for forming hierarchical groups of mutually exclusive subsets of variables. This R-mode cluster method identifies interrelationships between variables which are useful for variable selection and dimension reduction. Importantly, the method is based on metric elements defined on the sample space of variables. Consequently, hierarchical clustering of compositional parts should respect the particular geometry of the simplex. In this work, the connections between concepts such as distance, cluster representative, compositional biplot, and log-ratio basis are explored within the framework of the most popular R-mode agglomerative hierarchical clustering methods. The approach is illustrated in a paleoecological study to identify groups of species sharing similar behavior. [ABSTRACT FROM AUTHOR]

The area east of Varanasi is one of numerous places along the watershed of the Ganges River with groundwater concentrations of arsenic surpassing the maximum value of 10 parts per billion (ppb) recommended by the World Health Organization in drinking water. Here we apply geostatistics and compositional data analysis for the mapping of arsenic and iron to help in understanding the conditions leading to the occurrence of elevated level of arsenic in groundwater. The methodology allows for displaying concentrations of arsenic and iron as maps consistent with the limited information from 95 water wells across an area of approximately 210 km; visualization of the uncertainty associated with the sampling; and summary of the findings in the form of probability maps. For thousands of years, Varanasi has been on the erosional side in a meander of the river that is free of arsenic values above 10 ppb. Maps reveal two anomalies of high arsenic concentrations on the depositional side of the valley, which has started seeing urban development. The methodology using geostatistics combined with compositional data analysis is completely general, so this study could be used as a prototype for hydrochemistry mapping in other areas. [ABSTRACT FROM AUTHOR]

It often occurs in practice that it is sensible to give different weights to the variables involved in a multivariate data analysis—and the same holds for compositional data as multivariate observations carrying relative information. It can be convenient to apply weights to better accommodate differences in the quality of the measurements, the occurrence of zeros and missing values, or generally to highlight some specific features of compositional parts. The characterisation of compositional data as elements of a Bayes space, which is as a natural generalisation of the ordinary Aitchison geometry, enables the definition of a formal framework to implement weighting schemes for the parts of a composition. This is formally achieved by considering a reference measure in the Bayes space alternative to the common uniform measure via the well-known chain rule. Unweighted centred logratio (clr) coefficients and isometric logratio (ilr) coordinates then allow us to express compositions in real space equipped with (unweighted) Euclidean geometry. The resulting elements of real space generated by the clr coefficients or ilr coordinates are invariant to the scale of the original compositions, but the actual scale of the weights matters. In this work, these formal developments are presented and used to introduce a general approach for weighting parts in compositional data analysis. The practical use is demonstrated on simulated and real-world data sets in the context of the earth sciences. [ABSTRACT FROM AUTHOR]