I am looking for the Hurst exponent calculation methodology. Please suggest online materials / methodology papers.
Asked
Active
Viewed 3,162 times
4
-
Linking related question (more specific though): http://stats.stackexchange.com/questions/46553/assumptions-for-hurst-exponent-calculation – Andre Silva Nov 18 '15 at 14:27
-
You can find many implementations in Matlab in this link: http://prorum.com/index.php/2173/calcular-expoente-exponent-dependencia-dependence-temporais – DanielTheRocketMan Feb 09 '16 at 15:54
2 Answers
8
The calculation is covered on the related wikipedia page.
R has several implementations for this:
- The fArma package provides 10 different functions to estimate the Hurst exponent (see LrdModelling).
- The Rwave package has the hurst.est() function.
- The fractal package has the hurstACVF() function.
- The dvfBm package is intended entirely for this purpose: "Hurst exponent estimation of a fractional Brownian motion by using discrete variations methods in presence of outliers and/or an additive noise".
The methods covered by fArma are taken from the "Estimators for Long-Range Dependence: An Empirical Study" (Taqqu, Teverovsky, Willinger 1995).
Edit: Just to add from the M. Tibbits comment below, you can find the Hurst exponent code for Matlab offered under a BSD license. The description:
This is an implementation of the Hurst exponent calculation that is smaller, simpler, and quicker than most others. It does a dispersional analysis on the data and then uses Matlab's polyfit to estimate the Hurst exponent. It comes with a test driver that you can delete.

Shane
- 11,961
- 17
- 71
- 89
-
-
1No, I'm afraid not. But R is fully open source so all the algorithms above will be very transparent if you want to re-implement. – Shane Oct 08 '10 at 14:20
-
2
-
First result from google: [Hurst Exponential](http://www.mathworks.com/matlabcentral/fileexchange/9842) – M. Tibbits Oct 08 '10 at 14:59