Abstract

A joint conditional likelihood (JCL) method, which is a semiparametric approach, is proposed to estimate the parameters of a logistic regression model when two covariate vectors are missing separately or simultaneously. The proposed method uses one validation and three non validations data sets; it is an extension of the method of Wang et al. who studied the case of one covariate missing at random. The asymptotic results of the JCL estimators are established under the assumption that all covariate variables are categorical. Simulation results show that the proposed method is the most efficient compared to the complete-case, semi-parametric inverse probability weighting, and validation likelihood methods. The proposed methodology is illustrated by a real data example.

摘要

当两个协变量向量分别或同时缺失时，提出了一种联合条件似然（JCL）方法，它是一种半参数方法，用于估计逻辑回归模型的参数。所提出的方法使用一个验证和三个非验证数据集；它是 Wang 等人方法的扩展。谁研究了一个协变量随机缺失的情况。JCL 估计量的渐近结果是在所有协变量都是分类变量的假设下建立的。仿真结果表明，与完整案例、半参数逆概率加权和验证似然法相比，所提出的方法是最有效的。拟议的方法由一个真实的数据示例说明。