Conditional quantile estimation through optimal quantization

TL;DR

This paper introduces a nonparametric method for estimating conditional quantiles using optimal quantization, demonstrating its consistency and effectiveness compared to traditional methods through theoretical analysis and simulations.

Contribution

It proposes a novel quantization-based estimator for conditional quantiles, with proven consistency and convergence rates, advancing nonparametric quantile estimation techniques.

Findings

Estimator is consistent for fixed N

Approximation error decreases as N increases

Outperforms local constant/linear and nearest neighbor methods

Abstract

In this paper, we use quantization to construct a nonparametric estimator of conditional quantiles of a scalar response $Y$ given a d-dimensional vector of covariates $X$. First we focus on the population level and show how optimal quantization of $X$, which consists in discretizing $X$ by projecting it on an appropriate grid of $N$ points, allows to approximate conditional quantiles of $Y$ given $X$. We show that this is approximation is arbitrarily good as $N$ goes to infinity and provide a rate of convergence for the approximation error. Then we turn to the sample case and define an estimator of conditional quantiles based on quantization ideas. We prove that this estimator is consistent for its fixed-$N$ population counterpart. The results are illustrated on a numerical example. Dominance of our estimators over local constant/linear ones and nearest neighbor ones is demonstrated through extensive simulations in the companion paper Charlier et al.(2014b).