逐步回归的基本思想是在考虑的全部因素（变盘）中，按其对指标 YY作用的

The basic idea is that all factors stepwise regression (and variable) in consideration, according to their size on the role of indicators YY significant degree, descending introduced one by one regression equation, the equation contains all of the significant influence of variables, does not contain no significant effect of the variable. Each of introducing a factor (index), the need for a significant FF test, when the factor of influence was significant, which is incorporated, while each introduction of a new factor, the further need for the regression equations have been introduced factors one by one FF significance tests, such as influential found no significant factor, it is removed, until the regression equation factor no longer be excluded, and no new factors may be introduced, considered the end of the whole process.

The basic idea of step-by-step regression is to introduce regression equations from large to small, in terms of the magnitude of the effect on the indicator YY, in all factors considered (variants), so that the equations contain all the variables that have a significant effect on them, not the variables that have an insignificant effect on the index. Each factor (change of site) needs to be tested for a significant FF, when the factor has a significant effect, and, after each introduction of a new factor, a significant FF test is required for the factors already introduced in the regression equation, and if a factor with an unremarkable effect is found, it is rejected. The whole process is not completed until there is no rejectable factor in the regression equation and no new factor seditable.

The basic idea of stepwise regression is to introduce regression equation one by one from large to small in consideration of all factors (variable disk), according to the significance of its effect on the indicator YY, so that the equation contains all the variables that have significant influence, but not the variables that have no significant influence. For each factor (index) introduced, a significance FF test shall be conducted. When the influence of this factor is significant, it shall be introduced. At the same time, after the introduction of a new factor, the significance FF test shall be conducted for the factors already introduced in the regression equation one by one. If there is any factor with no significant impact, it shall be eliminated until there is no factor in the regression equation to be eliminated If there is no new factor to introduce, the whole process will be finished.