A comparison of some Hurst parameter estimators

In the last few years the long-range dependence analysis of time-series became more important. A key parameter characterizing long-range dependent processes is the Hurst parameter H. The goal of this paper is to compare some estimation techniques for the Hurst parameter. We found that the best estimator is the one based on the second order discrete wavelet transform statistical analysis and works for second order wide sense stationary random processes.