GROWTH-CURVES - A 2-STAGE NONPARAMETRIC APPROACH

被引：20

作者：

BOULARAN, J

FERRE, L

VIEU, P

机构：

[1] UNIV TOULOUSE 3,STAT & PROBABIL LAB,CNRS,URA 745,F-31062 TOULOUSE,FRANCE

[2] UNIV PAU & PAYS ADOUR,MATH APPL LAB,CNRS,URA 1204,F-64000 PAU,FRANCE

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 1994年 / 38卷 / 03期

关键词：

ADDITIVE MODEL; GROWTH CURVES; KERNEL SMOOTHING; NONPARAMETRIC MIXED EFFECTS MODEL; TWO-STAGE PROCEDURE;

D O I：

10.1016/0378-3758(94)90014-0

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

We develop in this paper a two-stage nonparametric method for estimating growth curves. Our two-stage additive model consists in assuming that dependence between height and age is a sum of two components, the former being the same for all the individuals. We estimate both functional components by nonparametric kernel smoothing techniques. We first give theoretical results concerning L1, L2 and L(infinity) rates of convergence for our estimates. Then we discuss how to choose the smoothing degree in an optimal way. This method is then applied to a real example.

引用

页码：327 / 350

页数：24

共 39 条

[1]

[Anonymous], 1988, EMPEC, DOI DOI 10.1007/BF01972448

[2] PRINCIPAL COMPONENTS-ANALYSIS OF SAMPLED FUNCTIONS [J].