Zerbe G O
Growth. 1979 Dec;43(4):263-72.
A new nonparametric method is proposed for the construction of percentile curves and normal ranges which can be used to classify an individual's growth, velocity, and acceleration of growth dynamically over a time interval selected for its biological importance. Although the procedure requires longitudinal data, it is not necessary that all subjects be measured at identical times. Unlike classic static methods it does not provide percentile curves that are typical of no one. The median curve is the actual curve of the most "central" individual. The method is applicable to growth curves of any form with a general computer program already available for many forms. The technique is demonstrated for growth in weight curves smoothed by high degree polynomials.
本文提出了一种新的非参数方法，用于构建百分位数曲线和正常范围，该方法可用于在因生物学重要性而选定的时间间隔内，动态地对个体的生长、生长速度和生长加速度进行分类。尽管该程序需要纵向数据，但不要求所有受试者在相同时间进行测量。与经典的静态方法不同，它所提供的百分位数曲线并非适用于所有人的典型曲线。中位数曲线是最“中心”个体的实际曲线。该方法适用于任何形式的生长曲线，并且已有适用于多种形式的通用计算机程序。本文通过对用高阶多项式平滑后的体重增长曲线演示了该技术。
