APPLICATION OF RANDOM WEIGHTING METHODS IN REGRESSION MODELS
随机加权法在线性模型中的应用

For the classical regression models the problem of simulation for the distribution func-tion of the student statistics of LS estimator of an estimable function is discussed under the as-sumption that the components of the error vector are independent but not identically distrib-uted and other assumptions.By means of asym ptotical expansion of corresponding statisticswe prove that the random weighting approximation can attain the accuracy of o (1/n~(1/2),similar results are obtained also for the problem of approximation for the distribution func-tion of the testing statistics t_(n-p).