Asymptotic consistency of median regression trees (Q5928933): Difference between revisions

Estimation of a function parameter by adaptive recursive partitioning of the covariate space is a well known and effective nonparametric statistical technique. A piecewise constant nonparametric estimate of a conditional median function based on a least absolute deviations regression tree has been proposed and discussed by \textit{L. Breiman} et al. [ Classification and regression trees, Wadsworth, Belmont, CA, (1984; Zbl 0541.62042)]. In this article, we derive and discuss some general regularity conditions that can ensure asymptotic consistency of such an estimate. We also discuss a weighted average technique following \textit{P. Chaudhuri} et al. [Stat. Sin. 4, No. 1, 143--167 (1994; Zbl 0824.62032)] based on smooth weight functions that can nicely glue the discontinuous constant pieces of such an estimate to yield a smooth and consistent estimate.