Flexible conditional density estimation for time series

TL;DR

FlexCodeTS is a versatile nonparametric conditional density estimator for time series that adapts its convergence rate based on the regression method used, outperforming some existing methods in empirical tests.

Contribution

Introduces FlexCodeTS, a flexible conditional density estimation method for time series that inherits the convergence rate of the underlying regression technique.

Findings

FlexCodeTS generally outperforms NNKCDE and GARCH in empirical tests.

FlexCodeTS adapts its convergence rate to the chosen regression method.

Empirical results show FlexCodeTS achieves the best performance in CDE and pinball loss.

Abstract

This paper introduces FlexCodeTS, a new conditional density estimator for time series. FlexCodeTS is a flexible nonparametric conditional density estimator, which can be based on an arbitrary regression method. It is shown that FlexCodeTS inherits the rate of convergence of the chosen regression method. Hence, FlexCodeTS can adapt its convergence by employing the regression method that best fits the structure of data. From an empirical perspective, FlexCodeTS is compared to NNKCDE and GARCH in both simulated and real data. FlexCodeTS is shown to generally obtain the best performance among the selected methods according to either the CDE loss or the pinball loss.