Schneiderman E D, Willis S M, Kowalski C J, Ten Have T R
Department of Oral and Maxillofacial Surgery, Baylor College of Dentistry, Dallas, TX 75246.
Comput Biol Med. 1993 Mar;23(2):149-54. doi: 10.1016/0010-4825(93)90146-r.
In several areas of biomedicine, one needs to predict future measurements for a growing individual on the basis of longitudinal data. Here we consider the problem of estimating the values of a given measurement for a particular individual at T-T* points in time, given T* observations on that individual, and all T values for a sample of N "similar" individuals. This extends our previous discussion [Schneiderman et al., Comput. Biol. Med. 22, 181-188 (1992)], which was limited to the case T* = T-1, to longer-term predictions. We again make a user-friendly GAUSS program available to perform the associated computations. Examples illustrating the use of the program and the accuracy of the predictions it provides are included.
在生物医学的多个领域，人们需要根据纵向数据预测成长中个体未来的测量值。在此，我们考虑这样一个问题：给定某个体的(T^)次观测值，以及(N)个“相似”个体样本的所有(T)次测量值，估计该特定个体在(T - T^)个时间点上给定测量值的值。这将我们之前的讨论[施奈德曼等人，《计算机生物学与医学》22，181 - 188（1992）]（该讨论仅限于(T^* = T - 1)的情况）扩展到了更长期的预测。我们再次提供一个用户友好的GAUSS程序来执行相关计算。文中包含了说明该程序使用方法及其所提供预测准确性的示例。
