Consider the mixed linear model Y = X Where Y is a random n X 1 Column vector.are known . are Parameters,This paper shows the maximum likehood estimate of en shows a kind ofminimax estimate forinhere fj. is a fixed effect, a, is a random effect.idiere , a, are red efecis a random effect.ahere is a fixed effect, aare random effects .inhere n is a fixed effect , are random effects .ttiere is a fixed effect, a, are random effects .inhere p is a fixed effect. a,are random effects, the maximum likehood estimates affixed effect are the least squares estimates in the above cases.Dear with the mixed linear medel, we consider the linear estimate choose a loss functiomuhere satisfy mm Let theorem 1 Junction Theorem).Let P = (PltP2),Q = (Qi, Q2), -P'l-Pi = Q'lQi = Imf, then the unique linear minimax estmate of SB is... |