"The overriding goal of this paper has been to move the analysis of incomplete data in the direction of modeling and away from the predominantly algorithmic approach that is commonly used. The framework for the modeling approach, including the use of the log-linear model, was originally suggested by Willekens (1982). This paper builds on his work by: (1) providing a justification for the use of synthetic data, (2) highlighting some of the key statistical qualities of the log-linear model in a GLM framework, and (3) proposing a path diagram to structure the analysis." The author includes an application to incomplete occupational migration and mortality tables for the United States. (EXCERPT)

中文翻译：

---

基于模型的不完整数据分析与职业流动和迁移账户的应用

“本文的首要目标是将不完整数据的分析转向建模方向，远离常用的主要算法方法。建模方法的框架，包括对数线性模型的使用，最初由 Willekens (1982) 提出。这篇论文建立在他的工作基础上：(1) 为使用合成数据提供了理由，(2) 突出了 GLM 框架中对数线性模型的一些关键统计特性，以及 (3) 提出一个路径图来构建分析。” 作者包括对美国不完整职业迁移和死亡率表的应用。（摘抄）