Key points are not available for this paper at this time.
We consider the performance of the bootstrap in high-dimensions for the setting of linear regression, where p < n but p/n is not close to zero. We consider ordinary least-squares as well as robust ...
Karoui et al. (Mon,) studied this question.