Abstract

Big Data are huge amounts of digital information that rarely result from properly planned surveys; as a consequence they often contain redundant observations. When the aim is to answer particular questions of interest, we suggest selecting a subsample of units that contains the majority of the information to achieve this goal. Selection methods driven by the theory of optimal design incorporate the inferential purposes and thus perform better than standard sampling schemes.

中文翻译：

---

大数据集的优化设计子采样

摘要

大数据是大量的数字信息，很少通过适当计划的调查产生；因此，它们通常包含多余的观察结果。当目的是回答感兴趣的特定问题时，我们建议选择包含大部分信息的单元子样本以实现此目标。由优化设计理论驱动的选择方法结合了推理目的，因此比标准抽样方案表现更好。