Hi, everybody.
My name is Ricardo and I'm doing doctorate at USP (Brazil). I have used a
lot the package 'fractal', developed by William Constantine and Donald
Percival, in my research.
However, I have a doubt about the execution of RoverS function. I was
reading an article (http://www.bearcave.com/misl/misl_tech/wavelets/hurst/)
about the hurst exponent and I resolved to compare the results present there
and the results provided by RoverS function. Initially, I got a synthetic
data set, available in
http://www.bearcave.com/misl/misl_tech/wavelets/hurst/brown72.h, and I
loaded the data set into the variable dataH and executed it using
RoverS(dataH) command.
So, this command returned the result:
> RoverS(dataH)
X
0.7165741
However, the result of the hurst exponent provided by the website is 0.7270.
Is it a problem or can I use the value 0.72? Can this difference cause
biggest errors in others data sets? Could anyone help me?
Thanks for your attention!
Ricardo
[[alternative HTML version deleted]]