How do you identify, measure, and mitigate multicollinearity in a high-dimensional dataset before feeding it into a linear model?
