Hello,
I'm using the fArma package to estimate Hurst exponent by R/S method.
I've
some measurements in the following format:
Call_number Call_duration
I'm using the following steps. Am new to R, so it would help if someone
could please confirm if my steps are correct.
Further, this method seems to give a value H ~= 0.8 (> 0.5). Is the use of
R/S method to estimate H OK?
> head(data, n=10)
X1 X5402
1 2 1876
2 3 7653
3 4 2969
4 5 6215
5 6 2596
6 7 2434
7 8 1380
8 9 937
9 10 2047
10 11 712>
> rsFit(data[,2])
Title:
Hurst Exponent from R/S Method
Call:
rsFit(x = data[, 2])
Method:
R/S Method
Hurst Exponent:
H beta
0.8030667 0.8030667
Hurst Exponent Diagnostic:
Estimate Std.Err t-value Pr(>|t|)
X 0.8030667 0.05039722 15.93474 4.643595e-17
Parameter Settings:
n levels minnpts cut.off1 cut.off2
227656 50 3 5 316
Description:
Sun May 6 22:07:43 2012 by user:
There were 14 warnings (use warnings() to see them)
--
Thanks,
Barun Saha
JPA
IIT, Kharagpur
http://pothi.com/pothi/book/barun-saha-swapner-kheya
http://delay-tolerant-networks.blogspot.com/p/one-tutorial.html
[[alternative HTML version deleted]]
