Each leave-one-out statistic has the same distribution as , soConsequently the bias of the jackknife bias correction isThe assumed expansion givesandTheir difference is therefore , as required.
Articles by others on the same topic
There are currently no matching articles.