ABSTRACT
Compositional data can be viewed as a non-negative vector with unit-sum constraint and therefore restricted to a simplex of appropriate dimension. A typical example of data of this nature would be compositional contents of alloys and ores. Analysis of such data also arises when dealing with synthesis of alloys or compounds i.e., in specifying a compositional range of contents to achieve a certain level of some desired physical property. In environmental studies, toxic contents in air reveal the extent of atmospheric pollution and analyzing these components helps in understanding the harmful effects of toxins on our health. In another case data on several life-forms in a particular region may be available as percentages and we may want to assess effect of time and other varying ecological conditions on their survival. Although not frequently identified as such, the scope for obtaining information compositionally is unlimited.
