UNIVERSALLY CONSISTENT CONDITIONAL U-STATISTICS

• 成果类型: Article
• 署名作者: STUTE, W
• 刊物名称: ANNALS OF STATISTICS
• ISSN/ISSBN: 0090-5364
• DOI: 10.1214/aos/1176325378
• 发表日期: 1994
• 页码: 460-473
• 关键词: regression

摘要

We introduce a general class of conditional U-statistics and present sufficient conditions for their universal consistency in rth mean. It is shown that under mild assumptions on the smoothing parameters, window and k(n)-nearest neighbor estimators are universally consistent. An application to a new nonparametric discrimination problem is also included.