This paoer deals witj the linear variande component model ILLMit is shown that a general result conecerning estimable funftuons in a general gauss markob model forms the foundation of minqe theory this is done by teansfornming linear midels generated by yy finally tne tsults are applied to case m=1 yielding new interesting fornulase for the bque of residual variance in the linear regression model
展开▼