Variance inflation factors may be reduced using ridge, surrogate and raised estimators.

What is it about?

This paper compares three methodologies to reduce collinearity in linear regression models. The three estimators are (i) ridge estimator of Hoerl; (ii) the surrogate estimator of Jensen and Ramirez and (iii) the raise estimator of Garcia.

Featured Image

Why is it important?

Collinearity in regression models leads to very unstable estimators where small changes in data may lead to large variations in the estimators.