Local bootstrap

✍ Scribed by Sheng G. Shi

• Book ID: 104629837
• Publisher: Springer Japan
• Year: 1991
• Tongue: English
• Weight: 466 KB
• Volume: 43
• Category: Article
• ISSN: 0020-3157
• DOI: 10.1007/bf00121646

✦ Synopsis

The local bootstrap method draws bootstrap samples from a neighborhood of each data point. Unlike the common bootstrap, it is especially useful for heteroscedastic data. This paper considers local bootstrap of y for a data set {(x~, y~)} with the neighborhood defined by the window in a kernel regression model. It shows that for each x, the bootstrap distribution and moments converge almost surely to the true distribution and moments of y respectively, and this convergence is also uniform in x within the data set.