在钢筋的研究中，我们希望从对因变量y有影响的诸多变量中选择一些变量作为

在钢筋的研究中，我们希望从对因变量y有影响的诸多变量中选择一些变量作为自变量，应用多元回归分析的方法建立“最优”回归方程，以便对因变量进行预报或控制，即在回归方程中包含所有对因变量y影响显著的自变量，而不包含对y影响不显著的自变量。逐步回归分析正是根据这种原则提出来的一种回归分析方法，它的主要思路是在全部考虑的自变量中按其对y的显著程度大小，由大到小地逐个引入回归方程，而对那些对y作用不显著的变量可能始终不被引入回归方程。另外，已被引入回归方程的变量在引入新变量后也可能失去重要性，需要从回归方程中剔除。在引入和剔除的过程中，每一步都要进行F检验，以保证在引人新变量前回归方程中只含有对y影响显著的变量，而不显著的变量已被剔除.

結果 (英文) 1:

In steel of research in the, we hope from on for variable y has effect of many variable in the select some variable as since variable, application multiple return analysis of method established "optimal" return equation, to on for variable for forecast or control, that in return equation in the contains all on for variable y effect significantly of since variable, and not contains on y effect not significantly of since variable. Stepwise regression analysis is based on the principle of a regression analysis, its key ideas were all considered significant levels of the variable y by its size, from largest to smallest by regression equations, and no significant role for those of y variable may not always be regression equations. In addition, variables have been introduced into the regression equation also may lose its importance after introduces new variables, need to be excluded from the regression equation. In the process of introducing and removing, every step is an f-test, to ensure that the introduction of new variables in regression equations containing only significant influence on the y variables without significant variables have been removed.

結果 (英文) 2:

In the study reinforced, we want to select some variables on the dependent variable y from many influential variables as independent variables, multiple regression analysis method to establish "optimal" regression equation for the dependent variable to predict or control, that contains all of the dependent variable y affect a significant independent variable in the regression equation, and does not contain no significant effect on the y argument. Stepwise regression analysis it is proposed according to a principle of such a regression analysis, its main idea is to consider all the arguments in its significant degree of y of size, from large to small regression equation introduced one by one, and for those who had no significant effect on the y variables may never be introduced into the regression equation. In addition, variable regression equation has been introduced after the introduction of new variables may lose importance, we need to eliminate from the regression equation. In the process of introducing and removing, the F-test every step to ensure that prior to the introduction of a new variable regression equation contains only y on significant variables, without significant variables have been removed.

結果 (英文) 3:

Bars in the study, we hope to choose some variables from many variables affect the dependent variable y as independent variables, multiple regression analysis method to establish the optimal regression equation, the dependent variables to forecast or control, which is included in the regression equation of dependent variable y significantly affected variables. But do not contain variables had no significant effect on y. Stepwise regression analysis is based on a regression analysis method put forward the principle, its main idea is to consider in all variables according to the significant degree on the size of the Y, from large to small individually into the regression equation, while those of Y had no significant effects on the variables may not always be introduced into the regression equation. In addition, the variables that have been introduced into the regression equation may lose their importance after introducing the new variables, which need to be eliminated from the regression equation. In the process of introduction and elimination, each step should be carried out F test to ensure that the new variables in the lead before the regression equation contains only significant variables affecting the Y, and not significant variables have been removed.