a. a linear transformation of a model does not change the model and should not change the estimates of a model. b. Since high intercorrelations among variables [can] simply be transformed into low variances on linear combinations of variables, multicollinearity is itself not a fundamental measure of the strength of the data. c. . ... ridge regression's ... weakness ... is the use of a loose representation of a priori beliefs and the reliance at times on ad hoc pseudoinformation.
No takes yet. Share an insight, caveat, or question.
Donald W. Marquardt (1980) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: