UNIFORM CONFIDENCE BOUNDS FOR NONPARAMETRIC REGRESSION
Steinar Bjerve, Kjell A. Doksum, Brian S. Yandell
Abstract
Open-access reader
Steinar Bjerve, Kjell A. Doksum, Brian S. Yandell
Abstract
Open-access reader
Let (X, Y) be a bivariate random variable and let m(x) = E(YIX = x) be the regression function of Y on X. Suppose that v 1 , .••,Yn are independent observations of Y at X= x 1 , .••,xn.We consider nearest neighbor estimates, m(x), and employ well-known inequalities to obtain exact and asymptotic uniform confidence bounds for Em(x) and m(x) based on m(x).Finally we discuss bias-properties of m(x).
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Let (X, Y) be a bivariate random variable and let m(x) = E(YIX = x) be the regression function of Y on X. Suppose that v 1 , .••,Yn are independent observations of Y at X= x 1 , .••,xn.We consider nearest neighbor estimates, m(x), and employ well-known inequalities to obtain exact and asymptotic uniform confidence bounds for Em(x) and m(x) based on m(x).Finally we discuss bias-properties of m(x).
Key concepts: Nonparametric statistics, Statistics, Regression, Nonparametric regression, Confidence interval, Regression analysis, Econometrics, Mathematics