While learning about regularization I came across this MATLAB article on regularization. It says
The penalty term is made more effective by using a positive definite matrix R, which allows weighting and/or rotation of the parameter vector:The square matrix R gives additional freedom for:
Shaping the penalty term to meet the required constraints, such as keeping the model stable
Adding known information about the model parameters, such as reliability of the individual parameters in the θ vector
I am hard pressed to find more information on R. Is there a recommended method on how to choose R to determine weights, keep the model stable or express reliability?
Any references for further study will be helpful.