Mathematical Sciences: Nonparametric Curve Estimation

数学科学：非参数曲线估计

• 批准号: 9203135
• 负责人: James Marron
• 金额: $ 16.83万
• 依托单位: University of North Carolina at Chapel Hill
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1992
• 资助国家: 美国
• 起止时间: 1992-08-01 至 1996-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9203135&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Nonparametric; Curve; Estimation

This project addresses the theory and application of smoothing methods in nonparametric curve estimation. The focus here is on bandwidth selection with exact risk calculation and qualitative smoothing and on an examination of the superiority of local polynomial smoothers to more commonly used methods. For local polynomial smoothers, this project will address issues of boundary effects, percentile regression and robustification, variable order of fit, goodness-of-fit testing and local weighted likelihood. Smoothing methods in nonparametric curve estimation constitute a flexible and powerful approach to data analysis which is of particular use in complicated situations where an exact or even useful model is unavailable. Such problems arise, for example, in economics, zoology, and marketing among other areas.Effective practical implementation of smoothing methods requires analytic understanding of their properties.

本课题主要研究了 非参数曲线估计中的平滑方法。重点 这里是关于带宽选择与精确的风险计算， 定性平滑和检查的优越性， 局部多项式平滑到更常用的方法。 为 本地多项式平滑器，该项目将解决的问题， 边界效应，百分位回归和鲁棒性， 变阶拟合、拟合优度检验和局部加权 可能性 非参数曲线估计中的光滑化方法 构成了一种灵活而强大的数据分析方法 这在复杂的情况下特别有用， 准确甚至有用的模型是不可用的。 出现了这样的问题， 例如，在经济学、动物学和市场营销等领域， 平滑方法的有效实际实施 需要对它们的性质进行分析理解。