非线性回归拟合正常人群前臂远端骨密度变化  

Fitting of bone mineral density changes in distal forearm of normal human by non-linear regression

作　　者：季颖[1] 华琳[2] 刘忠厚 

机构地区：[1]首都医科大学附属北京妇产医院,北京100006 [2]首都医科大学生物医学工程学院数学教研室 [3]中国老年学学会骨质疏松委员会

出　　处：《中国骨质疏松杂志》2005年第4期467-469,共3页Chinese Journal of Osteoporosis

摘　　要：目的采用非线性回归中的最大值函数拟合正常人群前臂远端骨密度(BMD)变化。方法用pDEXAOSTEOMETERDTX200骨密度仪测量603例5～89岁健康人前臂远端的BMD,对BMD随年龄的变化趋势采用非线性回归模型拟合并与线性回归模型作比较;根据最大值函数回归模型方程,以极大值作为峰值BMD计算出峰值BMD时相应的年龄值以及BMD下降13%和25%时相应的年龄值。结果以方程的决定系数R2来评价,非线性回归中最大值函数回归模型的拟合效果与线性回归模型中拟合效果最佳的三次回归模型相近。结论在BMD随年龄变化的研究中可试用非线性回归中的最大值函数进行拟合分析。Objective To fit the bone mineral density （BMD） changes in distal forearm of normal human by nonlinear regression. Methods BMDs of distal forearm were measured using pDEXA OSTEOMETER DTX-200 bone densitometer in 603 healthy subjects aged from 5 to 89 years old. Age-related changes were fitted by linear regression models and the maximal value regression model which was non-linear regression model. The ages of PBMD and the ages of 13 % and 25 % less than PBMD were calculated by maximal value regression model. Results The coefficients of determination （ R^2 ） for equations were similar in the maximal value regression model and the cubic regression model which was proved to be the best one in linear regression models. Conclusions The maximal value regression model may be considered as one of the methods in analyzing age-related changes of BMD.

关 键 词：骨密度 前臂 非线性回归 最大值回归模型 

分 类 号：R681[医药卫生—骨科学]