s*********e 发帖数: 1051 | 1 In a bagging application with large data set, bootstrap might be a bit
expensive in terms of computing cost. I am wondering what the potential (
negative) impact might be if jackknife instead of bootstrap is used.
if the answer is positive, what might be a reasonable proportion of training
data to be brawn in each iteration of jackknife sampling. Any related
experience or literature would be also highly appreciated.
thanks so much for your insight! | r****t 发帖数: 10904 | 2 Tu & Shao, Efron 1982 短一点,好读些 | s*********e 发帖数: 1051 | 3 十分感谢。能否把你说的文章寄给我一份?
l*******[email protected]
【在 r****t 的大作中提到】 : Tu & Shao, Efron 1982 短一点,好读些
|
|