I performed a discriminant factor analysis (DFA) between forest types and phenological data; I would like to constrain over the factors of the DFA, some biophisical drivers. Many thanks!
I recommend that you generate the discriminant analysis for your variables and technically support the redundancy of any variable with the support of the specific knowledge area.
In this way, you can have two perspectives, why choose or not a specific input, based on terms of computational statistics and from the point of view of the variable itself.