加些介绍
Why will it not suffice to screen data and remove outliers? There are several aspects to consider.
1. Users, even expert statisticians, do not always screen the data.
2. The sharp decision to keep or reject an observation is wasteful. We can do better by
down-weighting dubious observations than by rejecting them, although we may wish to
reject completely wrong observations.
3. It can be difficult or even impossible to spot outliers in multivariate or highly structured
data.
4. Rejecting outliers affects the distribution theory, which ought to be adjusted. In particular,
variances will be underestimated from the ‘cleaned’ data.
扫码加好友，拉您进群
全部版块
我的主页
收藏
