作者tren (窗外有蓝天)
看板Statistics
标题[问题] Logistic regression more robust to outliers than LDA
时间Sat Feb 10 04:49:48 2007
在 The Elements of Statistical Learning一书里,
作者於p105比较logistic regression(LR)与linear discriminant analysis(LDA)提到:
"...observations far from the decision boundary are down-weighted by
logistic regression..."
但小弟对於这里的down-weighting一直百思不解.由於LR和LDA有相同
的regression form,所以down-weighting应该不是因为log函数对
大数的compression而来.那麽,到底直观上或数学上,LR为何会对outliers
比较robust了?
谢谢!
--
※ 发信站: 批踢踢实业坊(ptt.cc)
◆ From: 168.122.81.155