Compositional Data Analysis (CoDA) refers to the analysis of all those vectors representing parts of a whole which only carry relative information. This type of data is ubiquitous in most applications, especially in geology, chemistry, genetics, and environmental sciences. The last comprehensive treatment was written by John Aitchison in 1986, and this book represents an update of that classic book.
Table of Contents
Interesting problems with compositional data. Constraints and problems they cause. Distance in simplex sample space. Covariance structure. Logistic normal and other distributions on the simplex. Log ratio analysis (alr, clr, ilr). Dimension reduction. Bases and compositions. Subcompositions and partitions. Zeros, rounding, measurement error. Compositions as covariates and mixtures. Integer compositions, multi-way compositions, compositional processes. Appendix A: R software for compositions. Appendix B: Bayesian analysis for compositional data.