Regression function estimation from dependent observations
We consider the problem of estimating a regression function with nonrandom design points and dependent errors. We construct a spline estimate of the regression function and obtain its rate of convergence. It turns out that the dependence of the observations is reflected in this rate.