Bayesian semi-parametric estimation of the long-memory parameter under FEXP-priors

TL;DR

This paper develops a Bayesian semi-parametric method using FEXP-priors for estimating the long-memory parameter in Gaussian time series, achieving near-minimax rates under certain prior choices.

Contribution

It introduces a Bayesian approach with specific priors for estimating the long-memory parameter, analyzing convergence rates and comparing to frequentist results.

Findings

Suboptimal convergence rates with certain priors.

Near-minimax rate achieved with sieve prior at specific rate.

Bayesian results align with frequentist estimators' performance.

Abstract

For a Gaussian time series with long-memory behavior, we use the FEXP-model for semi-parametric estimation of the long-memory parameter $d$. The true spectral density $f_o$ is assumed to have long-memory parameter $d_o$ and a FEXP-expansion of Sobolev-regularity $\be > 1$. We prove that when $k$ follows a Poisson or geometric prior, or a sieve prior increasing at rate $n^{\frac{1}{1+2\be}}$, $d$ converges to $d_o$ at a suboptimal rate. When the sieve prior increases at rate $n^{\frac{1}{2\be}}$ however, the minimax rate is almost obtained. Our results can be seen as a Bayesian equivalent of the result which Moulines and Soulier obtained for some frequentist estimators.