摘要
Consider the model (1.6), where eij (j=1,…,Ni, i=1,2,…) are i.i.d.with mean 0 andwriance 1.Introduce a randomly weighted estimate βn defined by (1.8).Assuming e11 ～N(0, 1)and,the paper gives a necessary and sufficient condition for βn to be a consistent estimateof β0,and under some further restrictions a normal approximation fo βn is established whichcan be used in constructing a large sample confidence interval of β0. Finally, in the non-normalcase a theorem about the consistency of βn is proved.
Consider the model (1.6), where eij (j=1,…,Ni, i=1,2,…) are i.i.d.with mean 0 andwriance 1.Introduce a randomly weighted estimate βn defined by (1.8).Assuming e11 ～N(0, 1)and,the paper gives a necessary and sufficient condition for βn to be a consistent estimateof β0,and under some further restrictions a normal approximation fo βn is established whichcan be used in constructing a large sample confidence interval of β0. Finally, in the non-normalcase a theorem about the consistency of βn is proved.
