局部多项式数据扩大法在群体药动学稀疏数据中的应用  

Data augmentation for sparse data in population pharmacokinetics based on local polynomial regression

作　　者：陶禹希[1] 陆涛[1,2] 刘俊麟[1] 言方荣[1,2] 王菲[1] 

机构地区：[1]中国药科大学基础部数学教研室,南京210009 [2]中国药科大学天然药物活性物质与功能国家重点实验室,南京210009

出　　处：《中国新药杂志》2013年第12期1361-1366,共6页Chinese Journal of New Drugs

基　　金：国家自然科学基金重点项目(NSFC 81130068);中央高校专项业务经费(JKQ2011032)

摘　　要：目的:针对药动学数据稀疏这一典型特征,讨论基于非参数局部多项式技术的数据扩大法在群体药动学稀疏数据中的应用。方法:介绍基于非参数局部多项式技术的数据扩大法的基本原理及方法步骤,通过计算机模拟技术分析该方法对参数估计的影响,并应用于头孢羟唑实际数据。结果与结论:与稀疏数据得到的参数估计相比,局部多项式数据扩大法所得参数估计更加准确可靠。局部多项式数据扩大法能够很好的解决群体药动学数据稀疏问题。Objective： Considering the typical character of data sparsity in pharmacokinetics, we proposed a method of data augmentation based on nonparametric local polynomial, and applied it to the sparse data in popu- lation pharmacokinetics study. Methods： We introduced the basic principle and implementation procedure of data augmentation method based on nonparametric local polynomial, and then studied the influence of this method on pa- rameter estimation via computer simulations. To illustrate the methodology, an application to the clinical data of ce- famandole was presented. Results and Conclusion： Compared with the parameters estimated through sparse data directly, the parameters estimated by local polynomial data augmentation method were more accurate and reasona- ble. Thus, local polynomial data augmentation method can be a good solution to the problem of sparse data in popu- lation pharmacokinetics.

关 键 词：稀疏数据 非参数估计 局部多项式 微分方程 群体药动学 

分 类 号：R911[医药卫生—药学]