Marginally specified logistic-normal models for longitudinal binary data
Marginally specified logistic-normal models for longitudinal binary data
复制标题
DOI：
10.1111/j.0006-341x.1999.00688.x
复制
发表时间：
1999-09-01
期刊：
影响因子：
1.9
通讯作者：
Heagerty, PJ
中科院分区：
文献类型：
--
作者：
Heagerty, PJ
鉴于计算方面的最新改进，可以使用广义线性混合模型获得纵向二进制数据的基于似然的推断（Breslow, N. 和 Clayton, D. G., 1993, 美国统计协会杂志 88, 9-25; Wolfinger, El. 和 O'Connell, M., 1993, 统计计算和模拟杂志 48, 233-243）接近。或者，Fitzmaurice 和 Laird (1993, Biometrika.80, 141-151)、Molenberghs 和。 Lesaffre (1994, Journal of the American Statistical Association 89, 633-644) 以及 Heagerty 和 Zeger (1996, Journal of the American Statistical Association 91, 1024-1036) 开发了一种基于似然的推断，该推断采用边际均值回归参数，并通过规范和/或边际高矩假设完成联合多元分布的完整规范。这些边际方法中的每一种都需要大量计算，并且目前仅限于较小的集群大小。在手稿中，采用了逻辑正态随机效应模型的替代参数化，并研究了参数估计的似然法和估计方程方法。所提出方法的一个关键特征是采用的边际回归参数仍然允许个体级别的预测或对比。给出了一个例子，其中科学兴趣在于重复测量之间的平均响应和协方差。
Likelihood-based inference for longitudinal binary data can be obtained using a generalized linear mixed model (Breslow, N. and Clayton, D. G., 1993, journal of the American Statistical Association 88, 9-25; Wolfinger, El. and O'Connell, M., 1993, Journal of Statistical Computation and Simulation 48, 233-243), given the recent improvements in computational approaches. Alternatively, Fitzmaurice and Laird (1993, Biometrika. 80, 141-151), Molenberghs and. Lesaffre (1994, Journal of the American Statistical Association 89, 633-644), and Heagerty and Zeger (1996, Journal of the American Statistical Association 91, 1024-1036) have developed a likelihood-based inference that adopts a marginal mean regression parameter and completes full specification of the joint multivariate distribution through either canonical and/or marginal higher moment assumptions. Each of these marginal approaches is computationally intense and currently limited to small cluster sizes. In the manuscript, an alternative parameterization of the logistic-normal random effects model is adopted, and both likelihood and estimating equation approaches to parameter estimation are studied. A key feature of the proposed approach is that marginal regression parameters are adopted that still permit individual-level predictions or contrasts. An example is presented where scientific interest is in both the mean response and the covariance among repeated measurements.