Font Size: a A A

Nonparametric confidence bands for regression curves

Posted on:2013-07-11Degree:Ph.DType:Dissertation
University:The University of Texas at DallasCandidate:Jin, SeoweonFull Text:PDF
GTID:1450390008487937Subject:Statistics
Abstract/Summary:
Nonparametric regression has gained much attention in recent years and has been widely applied to many fields such as biostatistics, econometrics, and engineering, due to its flexibility in modelling complex relationships among variables by the data themselves.;Finding a confidence band of an unknown regression curve is one of the most interesting aspects from parametric regression and nonparametric regression. In this dissertation, several well-known methods for finding confidence bands will be discussed, a nonparametric method using Hardle's method [Hardle, 1989] will be used for a confidence band of an unknown regression curve, and that method can be extended when responses are missing at random.;In chapter 1, I review some known linear regression results for confidence bands of unknown regression curves. In chapter 2, I review some known nonparametric regression results for confidence bands of unknown regression curves. In chapter 3, I find a confidence band of an unknown regression curve using Hardle's method [Hardle, 1989], a Nadaraya-Watson estimator, and a simulation result is included. In chapter 4, I find a confidence band of an unknown regression curve when responses are missing at random using the method in chapter 3, and a simulation result is included.
Keywords/Search Tags:Regression, Confidence bands, Nonparametric, Method, Chapter
Related items