Paragraph 37 contains incorrect _R_²
values for the lasso models. The corrected version reads:
The lasso (and best subset) models explained 69% (70%) of variation in lung cancer incidence, 55% (57%) in breast, 32% (34%) in colorectal, and 15% (19%) in prostate
More information is available from the corresponding GitHub issue.
…read more, vote or answer