2009 Workshop on Semiparametric methodology

2009年半参数方法研讨会

• 批准号: 0838278
• 负责人: Michael Daniels
• 金额: --
• 依托单位: University of Florida
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2008
• 资助国家: 美国
• 起止时间: 2008-11-01 至 2009-10-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0838278&HistoricalAwards=false
• 关键词: 2009; Workshop; Semiparametric; methodology

Semiparametric methods continue to be an active research area,especially as computing resources and power grow. These methods are usedfor inference in models with both parametric and nonparametriccomponents. Work in this setting ranges from semiparametric modeling andefficiency to semiparametric regression. The former involves makingefficient inferences in the presence of infinite dimensional nuisanceparameters. The latter corresponds to regression models with parametricerrors and unspecified (nonparametric) regression functions. Fundamentalwork remains to be done in high dimensional missing data problems, bothin the setting of frequentist and Bayesian inference and in estimatingregression functions with the complications posed in many differentareas of application. The workshop brings together leaders in the fieldof semiparametric methodology and young researchers. The invitedspeakers all are very distinguished individuals who have doneoutstanding work in semiparametrics and have co-authored three of theseminal texts. Flexible modeling with few assumptions is an integral component ofsolving many scientific problems. Semiparametric methods are a way toaccomplish this. Major advances have been made in the past thirty years.However, fundamental work remains to be done in high dimensionalproblems with incomplete data, in estimating regression functions withthe complications posed in many different areas of application, amongother areas. Applications for such work are wide ranging and includespatio-temporal models, capture-recapture models, and measurement errormodels in environmental applications, classification and testing in ingenetics and genomics, and missing data models in longitudinalobservational studies and clinical trials. The workshop provides anexcellent opportunity to discuss the many recent significantdevelopments in semiparametric methodology and to identify importantproblems and new research directions.

半参数方法仍然是一个活跃的研究领域，特别是随着计算资源和能力的增长。这些方法用于在具有参数和非参数组件的模型中进行推理。此设置中的工作范围从半参数建模和效率到半参数回归。前者涉及在存在无限维干扰参数的情况下进行有效的推理。后者对应于具有参数误差和未指定（非参数）回归函数的回归模型。在高维缺失数据问题上，无论是在频率论和贝叶斯推理的设置中，还是在估计回归函数以及在许多不同应用领域中提出的复杂性方面，仍有基础工作要做。该研讨会汇集了半参数方法领域的领导者和年轻研究人员。受邀的演讲者都是非常杰出的人士，他们在半参数学领域做出了杰出的工作，并与人合着了三篇重要的文章。几乎没有假设的灵活建模是解决许多科学问题的一个组成部分。半参数方法是实现这一目标的一种方法。在过去的三十年中已经取得了重大进展。然而，在具有不完整数据的高维问题、在估计回归函数以及在许多不同应用领域等领域中出现的复杂性方面仍有基础工作要做。此类工作的应用范围很广，包括环境应用中的时空模型、捕获-再捕获模型和测量误差模型，遗传学和基因组学中的分类和测试，以及纵向观察研究和临床试验中的缺失数据模型。该研讨会提供了一个绝佳的机会来讨论半参数方法论的许多最新重大发展，并确定重要问题和新的研究方向。