User:Attilauva/sandbox

Redundancy analysis (RDA) allows the user to derive a specified number of synthetic variables from one set of (explanatory) variables that explain as much variance as possible in another (response) set. It is a multivariate analogue of regression.

If we have a set of explanatory variables X = (X1, ..., Xn) and a set of response variables Y = (Y1, ..., Ym), and there are correlations among the variables, then RDA will find linear combinations of the X variables which have maximum correlation with all the Y variables.