Department of Botany, Stockholm University, SE-106 91 Stockholm, Sweden.
Ecol Lett. 2010 May;13(5):E7-9. doi: 10.1111/j.1461-0248.2010.01460.x.
Murtaugh (2009) recently illustrated that all subsets variable selection is very similar to stepwise regression. This, however, does not necessarily mean both methods are useful. On the contrary, the same problems with overfitting should apply. Ecologists should, if model building is indeed necessary, consider more reliable regression methods now available.
莫尔塔夫（2009）最近说明了所有子集变量选择非常类似于逐步回归。然而，这并不一定意味着这两种方法都有用。相反，过拟合的相同问题也应该适用。如果模型构建确实有必要，生态学家现在应该考虑更可靠的回归方法。
