问题的提出应用多元回归和逐步回归的前提是有关自变量必须是相互独立的。但是,事实上,在医学领域内各变量相关的情况大量存在。例如,身高、体重、肺活量、心象面积等,若以身高为依变量,则另外三个变量的独立性不能得到满足,也就是具有多元共线性。这时,其相关矩阵是病态的,进行矩阵初等变换后近于奇异矩阵,且回归 The premise of the application of multivariate regression and stepwise regression in the proposed problem is that the relevant independent variables must be independent of each other. However, in fact, there are a large number of variables-related situations in the medical field. For example, height, weight, vital capacity, and mental image area, if height is a dependent variable, the independence of the other three variables cannot be satisfied, that is, multivariate collinearity. At this time, the correlation matrix is ​​ill-conditioned. After the elementary transformation of the matrix, the matrix is ​​close to the singular matrix and the regression is performed.