文献类型：期刊文献

英文题名：Parameter estimation of nonlinear mixed-effects models using first-order conditional linearization and the EM algorithm

作者：Fu, Liyong[1] Lei, Yuancai[1] Sharma, Ram P.[2] Tang, Shouzheng[1]

第一作者：符利勇

通信作者：Tang, SZ[1]

机构：[1]Chinese Acad Forestry, Res Inst Forest Resource Informat Tech, Beijing, Peoples R China;[2]Norwegian Univ Life Sci, Dept Ecol & Nat Resource Management, As, Norway

年份：2013

卷号：40

期号：2

起止页码：252-265

外文期刊名：JOURNAL OF APPLIED STATISTICS

收录：;Scopus(收录号：2-s2.0-84871189546);WOS:【SCI-EXPANDED(收录号:WOS:000312341300002)】；

基金：The authors are very grateful to Dr. Mingliang Wang, Dr. Xiangdong Lei, the Associate Editor, and two anonymous referees for valuable comments on an earlier draft of the manuscript. The authors also express their appreciation to the Chinese National Natural Science Foundation (NO. 31170588) and the Forestry Public Welfare Scientific Research Project (No. 201004002) for financial support for this study, and the Research Institute of Forestry of the Chinese Academy of Forestry for providing the Chinese fir data.

语种：英文

外文关键词：Cunninghamia lanceolata; expectation-maximization algorithm; first-order conditional expansion; Lindstrom and Bates algorithm; nonlinear mixed-effects models; orange tree data; simulated data

摘要：Nonlinear mixed-effects (NLME) models are flexible enough to handle repeated-measures data from various disciplines. In this article, we propose both maximum-likelihood and restricted maximum-likelihood estimations of NLME models using first-order conditional expansion (FOCE) and the expectationmaximization (EM) algorithm. The FOCE-EM algorithm implemented in the ForStat procedure SNLME is compared with the Lindstrom and Bates (LB) algorithm implemented in both the SAS macro NLINMIX and the S-Plus/R function nlme in terms of computational efficiency and statistical properties. Two realworld data sets an orange tree data set and a Chinese fir (Cunninghamia lanceolata) data set, and a simulated data set were used for evaluation. FOCE-EM converged for all mixed models derived from the base model in the two realworld cases, while LB did not, especially for the models in which random effects are simultaneously considered in several parameters to account for between-subject variation. However, both algorithms had identical estimated parameters and fit statistics for the converged models. We therefore recommend using FOCE-EM in NLME models, particularly when convergence is a concern in model selection.