Mathematical Sciences: Computationally Intensive Problems in Statistics

数学科学：统计中的计算密集型问题

• 批准号: 8916245
• 负责人: Naomi Altman
• 金额: --
• 依托单位: Cornell Univ - State: AWDS MADE PRIOR MAY 2010
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1990
• 资助国家: 美国
• 起止时间: 1990-03-01 至 1992-02-29
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8916245&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Computationally; Intensive; Problems

This research considers computationally intensive methods in statistics that involve the use of high speed workstations. One problem involves the combined application of data-analytic methods of nonparametric regression and empirical Bayes techniques to improved fitting and prediction for growth curve data. Another problem involves mixed models incorporating both fixed and random effects for binary data. In this problem the focus is on variance components estimation in mixed models and the analogues of best linear unbiased prediction of the observed values of the random effects. These will include models for the repeated measures and longitudinal data settings.

这项研究认为，计算密集型的统计方法，涉及使用高速工作站。其中一个问题涉及非参数回归和经验贝叶斯技术的数据分析方法的组合应用，以改进生长曲线数据的拟合和预测。另一个问题涉及混合模型，包括二进制数据的固定和随机效应。在这个问题中，重点是在混合模型的方差分量估计和随机效应的观测值的最佳线性无偏预测的类似。这些将包括重复测量和纵向数据设置的模型。