Barrett_Zh
大人群筛查资料,想要了解发病率情况,缺失40%左右,统计学上有好的补救方法吗?
ilikemath
non-ignorable missing,you can use semi-parametric approach.
wxq1638
应该瘪了
rtist
an important question is why so many data are missing. Without knowing this, it's hard to rescue.
yihui
统计本身又不是救世主,还好不是缺90%……虽然缺失数据有插补方法,但是一定要知道缺失的机理(MAR?MCAR?等等……)
楼主请修改你的帖子标题!看看论坛说明以及别人是怎么提问的。