摘要
设M_1和M_2是两个带有预测量的线性模型,通过使用矩阵秩方法,本文给出了模型M_1下预测量的最优线性无偏预测同时也是模型M_2下的最优线性无偏预测的充分必要条件.作为这个结果的应用,我们给出了两个线性混合模型间最优线性无偏预测等价性的充分必要条件.
Let M1 and M2 be two linear models with new observations. Through matrix rank method, we derive the necessary and sufficient conditions for the best linear unbiased predictor （BLUP） of the new observation under the model M1 is also BLUP under the model M2. As applications, the conditions of equality of the BLUPs under two mixed linear models are also given.
出处
《应用概率统计》
CSCD
北大核心
2011年第5期522-532,共11页
Chinese Journal of Applied Probability and Statistics
基金
supported in part by Shanghai Municipal Natural Science Foundation(10ZR1420600)
Foundation of Shanghai Municipal Education Commission(11ZZ182)
关键词
一般线性模型
混合线性模型
最优线性无偏预测
矩阵秩方法
General linear model, mixed linear model, best linear unbiased predictor （BLUP）, matrix rank method.
