纵向单调缺失数据下线性模型的二次推断函数估计
QUADRATIC INFERENCE FUNCTIONS ESTIMATION IN LINEAR MODELS FOR LONGITUDINAL DATA WITH MONOTONE MISSING PATTERNS
纵向数据缺失的情况常常发生, 文章考虑响应变量带有单调缺失的纵向数据. 在逆概率加权广义估计方程~(IPWGEE)~的基础上, 采用二次推断函数~(QIF)~方法研究线性模型下回归参数的估计问题. 在一定的正则条件下, 证明了所得估计量的相合性和渐近正态性. 最后, 通过模拟研究和实例分析验证了所提出方法在有限样本下的实际表现.
In longitudinal studies, missing observations occur commonly. In this paper, we develop the method using quadratic inference function~(QIF) to handle longitudinal data with monotone missing observations present in response variables based on inverse probability-weighted generalized estimating equations~(IPWGEE). Under some appropriate conditions, we show that the QIF estimator is consistent and asymptotically normal. Finally, finite sample performance is assessed through simulation studies and a real data example.
纵向数据 / 单调缺失 / 逆概率加权广义估计方程 / 二次推断函数 / 线性模型. {{custom_keyword}} /
/
〈 | 〉 |