论文信息 - The BLUPs are not "best" when it comes to bootstrapping

The BLUPs are not "best" when it comes to bootstrapping

In the setting of mixed models, some researchers may construct a semiparametric bootstrap by sampling from the best linear unbiased predictor residuals. This paper demonstrates both mathematically and by simulation that such a bootstrap will consistently underestimate the variation in the data in finite samples.

Jeffrey S. Morris

[1] B. Efron,et al. The Jackknife: The Bootstrap and Other Resampling Plans. , 1983 .

[2] B. Efron. The jackknife, the bootstrap, and other resampling plans , 1987 .

[3] Robin Thompson,et al. Estimation of genetic parameters. , 2005 .

[4] Terry Speed,et al. [That BLUP is a Good Thing: The Estimation of Random Effects]: Comment , 1991 .

[5] D. Bates,et al. Mixed-Effects Models in S and S-PLUS , 2001 .

[6] G. Robinson. That BLUP is a Good Thing: The Estimation of Random Effects , 1991 .

[7] C. Mcgilchrist,et al. The derivation of blup, ML, REML estimation methods for generalised linear mixed models , 1995 .

[8] A. Rao,et al. Estimation of Genetic Parameters: principles , 2003 .