Show that we can obtain the ridge regression estimator from ordinary least squares on an augmented data set where x is centered and augmented with a k×k identity matrix, and y is augmented with a k×1 zero vector?