论文部分内容阅读
目的确定小儿手术麻醉期间机械通气时潮气量的影响因素。方法采用ＳＡＳ软件对资料进行残差分析，共线诊断并利用原点迫近回归实现岭回归分析。结果在残差图上，学生化残差大致落在｜ｒ｜≤２的水平带状区域内且无任何趋势。一般回归分析的标准误大多大于１．５；身高和体重的方差构成比分别为９０．８７％和５１．３９％，均超过了５０％；在岭迹图２中，当ｋ＝０．３时，岭迹大体稳定，据此所拟合的岭回归方程的所有回归系数标准误均小于０．１。结论残差分析表明本资料近似满足误差方差齐性假设，模型正态性假设；但由于回归系数标准误偏大，因此利用一般回归分析方法所建立的回归方程系数不够稳定，加之用诊断检验查出身高和体重间存在多重共线性，故本资料用一般回归分析方法拟合并不适合；采用岭估计重新分析，不仅减小了多元共线性的效应，而且降低了回归系数的标准误，获得了较为可靠的结果：确认体重、年龄、平均动脉压对潮气量起重要作用。
Objective To determine the influencing factors of tidal volume during mechanical ventilation during pediatric anesthesia. Methods The SAS software was used to analyze the residual data, diagnose the collinear problem and make use of the origin of the approaching regression to achieve ridge regression analysis. Results In the residual plot, the residual of studentization falls within the horizontal band region of | r | ≤2 without any trend. The standard error of general regression analysis was mostly greater than 1.5; the variance ratio of height and weight were 90.87% and 51.39% respectively, both of which exceeded 50%; in ridge trace 2, when k = 0.3 , The ridge trace is generally stable, and the standard errors of all the regression coefficients of the ridge regression equations fit are less than 0.1. Conclusion Residual analysis shows that this data approximates the hypothesis of homogeneity of error variance and model normality assumption. However, due to the large standard error of regression coefficient, the coefficient of regression equation established by general regression analysis method is not stable enough, There is multiple collinearity between height and weight, so this data is not fitted by general regression analysis method. Using ridge estimate reanalysis not only reduces the effect of multivariate collinearity, but also reduces the standard error of regression coefficient to obtain The more reliable results: confirm body weight, age, mean arterial pressure on tidal volume play an important role.