相对权重法在线性模型自变量相对重要性中的估计及其应用  

The Estimation and Application of Relative Weight of Relative Importance of the Independent Variable in Linear Model

作　　者：代鲁燕[1] 沈其君[1] 张波[1] 黄启风[1] 

机构地区：[1]宁波大学医学院预防医学系,315211

出　　处：《中国卫生统计》2013年第1期19-20,22,共3页Chinese Journal of Health Statistics

基　　金：国家自然科学基金(81172771)

摘　　要：目的在多元线性回归模型中,估计自变量的相对重要性。方法运用相对权重法估计各变量的相对重要性,并应用于肝手术病人预计存活时间影响因素的评价。结果血凝素、预后指数、酶功能的相对权重分别为:0.142、0.341、0.489;各变量对存活时间的贡献比例分别为:14.6%、35.1%、50.3%。结论酶功能对肝手术病人预计存活时间的影响最大,其次为预后指数,血凝素最小。当自变量间存在相关时,相对权重法估计的自变量相对重要性结果更稳定精确,更加符合实际情况。Objective Estimating the importance of each variable in multiple linear regression model. Methods Using the method of relative weight to estimate the importance of each variable and appling to evaluation the factors of survival time. Results The relative weights of hemagglutinin, prognostic index, enzyme function are respectively 0. 142, 0. 341,0. 489 ;the proportionate contribution of each variable makes to survival time is respectively 14. 6%, 35.1%, 50.3 %. Conclusion The maximum enzyme function of survival time, followed by prognostic index, the minimum hemagglutinin. When the correlation between independent variables, the method of relative weight estimating the importance of each variable is more stable, accurate and more realistic situation.

关 键 词：多元线性模型 相对权重 相对重要性 

分 类 号：O212.1[理学—概率论与数理统计]