Mathematical Sciences: Statistical Procedures for Errors in Variables Regression and other Multivariate Models

数学科学：变量回归和其他多元模型中误差的统计过程

• 批准号: 9002847
• 负责人: Leon Gleser
• 金额: $ 8.31万
• 依托单位: University of Pittsburgh
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1990
• 资助国家: 美国
• 起止时间: 1990-06-01 至 1992-11-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9002847&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Statistical; Procedures; Errors

This research will focus on topics in statistical theory and will be aimed at developing procedures in the areas of linear errors-in-variables regression models and goodness of fit procedures. The project is in the general area of statistics and probability. Errors-in-variables regression models describe relationships among variables when both the explanatory and response variables are subject to random error. Emphasis in this research will be placed on the role of reliability of measurements in determining the performance of errors-in-variables regression techniques. Goodness of fit procedures assess the adequacy of distributional assumptions or of structural models. This research will involve a study of general multinomial-based procedures for testing fit in categorical data.

这项研究将集中在统计理论的主题，并将致力于开发线性误差变量回归模型和拟合优度程序领域的程序。这个项目是在统计和概率的一般领域。变量误差回归模型描述了当解释变量和响应变量都存在随机误差时变量之间的关系。本研究的重点将放在测量的可靠性在确定变量误差回归技术的性能中的作用。拟合优度程序评估分布假设或结构模型的充分性。这项研究将涉及一般的基于多项式的程序的研究，以测试拟合在分类数据。