Bayesian Evidence Synthesis: Exploring the Limits of Limited data

贝叶斯证据合成：探索有限数据的局限性

• 批准号: RGPIN-2019-03957
• 负责人: Gustafson, Paul
• 金额: $ 3.86万
• 依托单位: University of British Columbia
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=737276
• 关键词: Bayesian; Evidence; Synthesis; Exploring; Limits

The quantification of evidence contained in data is clearly central to most, if not all, scientific endeavours. Very often the available data are more limited than one would like, yet the need to quantify the evidence is no less pressing. A lot of statistical methodological research, including that proposed here, looks to develop methodologies which mitigate limitations of the data. Common examples of the limitations faced include important variables being measured poorly, or not at all, for some study units, and bias in the selection of study units. This proposed research work aims to further the understanding and implementation of Bayesian methods to quantify evidence from data. The Bayesian paradigm offers a principled scheme to manage the various uncertainties that lie on the route from data to scientific evidence.

数据中所含证据的量化显然是大多数（如果不是全部的话）科学努力的核心。现有的数据往往比人们所希望的更为有限，但对证据进行量化的必要性同样紧迫。许多统计方法学研究，包括这里提出的，都希望开发出减轻数据局限性的方法。所面临的限制的常见例子包括一些研究单位的重要变量测量得很差，或根本没有测量，以及在选择研究单位时存在偏见。这项拟议的研究工作旨在进一步理解和实施贝叶斯方法来量化数据中的证据。贝叶斯范式提供了一个原则性的方案来管理从数据到科学证据的过程中存在的各种不确定性。