The Balanced LS Estimation of the Regressive Coefficient in Linear Model and Its Prediction;

With this new standard,we have obtained the generalized balanced LS estimation of the regression coefficients,and analyzed both the balanced LS estimation and generalized balanced LS estimation when the parameters are restricted by linear equations and boundaries.

The relationships between LS estimates of parameters under perturbed model and those under original model are given,and the equivalence of casedeletion model and meanshift model in parameter estimation for general covariane analysis model is illustrated.

The law of the variable selection on forest volume estimation is analyzed in detail by means of the real example, at the same time the predicting potentiality of forest volume through LS estimation is also discussed.

In this paper, another calculation on the weights of combination forecasting is proposed, that is, from the view of statistics, weight s LS estimation and its test of significance are given, in order to improve combination forecasting accuracy.

The least square estimation (LS estimation) theory is built on the basis that the error is normally distributed and the function model is linear.

According to the theory of Least-Square estimation,we analyze the problems existing in parameter estimation method for the Cobb-Douglas function model.

According to the theory of Least-Square estimation,we analyze the problem of parameter estimation precision existing in linearization model of nonlinear regression.