Wavelet threshold estimators for data with stationary correlated noise are constructed by applying a level-dependent soft threshold to the coefficients in the wavelet transform. A variety of threshold choices is proposed, including one based on an unbiased estimate of mean-squared error. The practical performance of the method is demonstrated on examples, including data from a neurophysiological context. The theoretical properties of the estimators are investigated by comparing them with an ideal but unattainable `bench-mark', that can be considered in the wavelet context as the risk obtained by ideal spatial adaptivity, and more generally is obtained by the use of an `oracle' that provides information that is not actually available in the data. It is shown that the level-dependent threshold estimator performs well relative to the bench-mark risk, and that its minimax behaviour cannot be improved on in order of magnitude by any other estimator. The wavelet domain structure of both short- and long-range dependent noise is considered, and in both cases it is shown that the estimators have near optimal behaviour simultaneously in a wide range of function classes, adapting automatically to the regularity properties of the underlying model. The proofs of the main results are obtained by considering a more general multivariate normal decision theoretic problem.

Wavelet threshold estimators for data with correlated noise

Estimating volatility from recent high frequency data, we revisit the question of the smoothness of the volatility process. Our main result is that log-volatility behaves essentially as a fractional Brownian motion with Hurst exponent H of order 0.1, at any reasonable timescale. This leads us to adopt the fractional stochastic volatility (FSV) model of Comte and Renault [Long memory in continuous-time stochastic volatility models. Math. Finance, 1998, 8(4), 291–323]. We call our model Rough FSV (RFSV) to underline that, in contrast to FSV, . We demonstrate that our RFSV model is remarkably consistent with financial time series data; one application is that it enables us to obtain improved forecasts of realized volatility. Furthermore, we find that although volatility is not a long memory process in the RFSV model, classical statistical procedures aiming at detecting volatility persistence tend to conclude the presence of long memory in data generated from it. This sheds light on why long memory of volatil...

Volatility is rough

We study the convergence of a generalisation of the quadratic variations of a Gaussian process. We build a convergent estimator of the local Holder index of the sample paths and prove a central limit theorem.

/pdf/quadratic-variations-and-estimation-of-the-local-holder-3xutxrwhrf.pdf

Quadratic variations and estimation of the local Hölder index of a gaussian process

Estimating volatility from recent high frequency data, we revisit the question of the smoothness of the volatility process. Our main result is that log-volatility behaves essentially as a fractional Brownian motion with Hurst exponent H of order 0.1, at any reasonable time scale. This leads us to adopt the fractional stochastic volatility (FSV) model of Comte and Renault. We call our model Rough FSV (RFSV) to underline that, in contrast to FSV, H<1/2. We demonstrate that our RFSV model is remarkably consistent with financial time series data; one application is that it enables us to obtain improved forecasts of realized volatility. Furthermore, we find that although volatility is not long memory in the RFSV model, classical statistical procedures aiming at detecting volatility persistence tend to conclude the presence of long memory in data generated from it. This sheds light on why long memory of volatility has been widely accepted as a stylized fact. Finally, we provide a quantitative market microstructure-based foundation for our findings, relating the roughness of volatility to high frequency trading and order splitting.

Fractional Gaussian noise, functional MRI and Alzheimer's disease.

The wavelet transform of Gaussian stationary continuoustime processes is studied. At each resolution, two Gaussian discrete stationary processes are obtained: the sequence of the approximations and the sequence of the details. Using the sequence of approximations of a continuous-time process, an approximation of the original one is reconstructed. We give the link between these sequences and the original process. Then we obtain the exact error between the approximation and the process, and the rate of convergence of this error. The weak convergence of the approximated process to the original process is proved. The asymptotic behaviour of the variance of the details and of the covariance between the details and the approximations are explained. This induced a choice of a best wavelet basis for data compression.

Jacques Istas

Papers

Quadratic variations and estimation of the local Hölder index of a gaussian process

Wavelet coefficients of a gaussian process and applications

Variations quadratiques et estimation de l'exposant de holder local d'un processus gaussien