论文部分内容阅读
通径系数实际上就是标准偏回归系数.在多元回归分析时,有一个在实际上经常遇到的问题,就是如何判别在所考虑的一些因素中,哪几个是影响因变数y的主要因素,哪几个是次要因素.为了分清k个自变数对因变数y影响的主次关系,一个自然的想法就是比较各个自变数的回归系数b_1、b_2、……b_k的绝对值大小.但是不能把回归系数直接比较,因为各个因变数和自变数所取单位不同,量纲不同,就没有直接比较的共同基础.为了消除这种单位不同的影响,人们采用了标准偏回归系数的做法.这也是通径分析的基础.
Path coefficient is actually the standard partial regression coefficient in the multiple regression analysis, there is a problem often encountered in practice, that is, how to determine the number of factors under consideration, which is the main factor affecting the dependent variable y , Which is the secondary factor.To distinguish the primary and secondary effects of k independent variables on the dependent variable y, a natural idea is to compare the absolute values of the regression coefficients b_1, b_2, ... b_k of each independent variable, We can not directly compare the regression coefficients, because each of the dependent variables and independent variables takes different units and different dimensions, so there is no common basis for direct comparison. In order to eliminate the different effects of such units, the standard partial regression coefficient is adopted. This is also the basis for path analysis.