Average squared residuals approach for testing linear hypotheses in nonparametric regression

What is it about?

We construct linear hypotheses tests for the regression function in nonparametric regression model in the case of a homoscedastic error structure and a fixed design. The test statistic we use is based on the average of squared residuals. We show the asymptotic normality of this statistic under the null hypothesis and the alternative. The work also contains a simulation study to investigate the finite sample behavior of the proposed test.

Featured Image