Double Normalization Algorithm for Ranking Individuals
Abstract
There are situations in which the magnitude of the variables is not always known. In these cases, and for certain problems, it is essential to understand the distribution of a set of variables per individual and simultaneously normalise each variable according to the position of each individual. Thus, the ranking of individuals within each variable, in relation to the distribution of individuals across a set of variables, can be generally described by an algorithmic approach that we shown as the Double Normalisation Algorithm for Ranking Individuals.
This algorithm was applied to the problem of identifying profiles in relation to the human microbiome and produced very promising results.