重复观测数据的半参数回归分析  

Semiparametric Regression Analysis of Repeated Measurements

作　　者：王文昌[1] 易东[1] 张蔚[1] 许汝福[1] 李辉智[2] 

机构地区：[1]第三军医大学卫生统计学教研室,400038 [2]西南政法大学刑侦系,400031

出　　处：《中国卫生统计》1999年第4期200-202,共3页Chinese Journal of Health Statistics

基　　金：国家自然科学基金!39870691

摘　　要：目的研究重复观测数据的回归分析技术。方法利用半参数回归分析的原理与方法，结合重复观测数据的特点，建立重复观测数据的半参数回归模型．并进一步讨论模型参数的估计方法及假设检验公式。结果讨论了重复观测数据的半参数回归模型的模型误差，分析了重复因素的效应及参数的影响，给出了其模型的方差分析表。结论通过实例分析，表明对重复观测数据的处理，半参数回归分析的效果优于普通的最小二乘法和广义最小二乘法。Objective To set up the semiparametric regression models on repeted measurements．Methods Using the theories and methods of semiparametric regression to establisn tne regression model of repeated measurements and discuss its methed of parametric estimation and hypothesis testing． Results The table of the analysis of variance about this model has been got, this table can be used to analyze the effect of the repeated factors and the paramers. Conclusion By analyzig one example, the result shows that semipeetric regression model is better than the ordinary least squares and the generalized least square analysis.

关 键 词：重复测量数据 半参数回归 参数估计 假设检验 

分 类 号：R195[医药卫生—卫生统计学]