[R-SIG-Finance] reliable Hurst exponent estimation

Martin Maechler maechler at stat.math.ethz.ch
Fri Aug 8 19:02:11 CEST 2008

>>>>> "tiu" == tolga i uzuner <tolga.i.uzuner at jpmorgan.com>
>>>>>     on Fri, 8 Aug 2008 13:18:20 +0100 writes:

    tiu> Dear R Finance Users, I am trying to reliably estimate
    tiu> the Hurst exponent using the ***fit functions in
    tiu> fArma. The estimates appear to be substantially
    tiu> different depending on the estimator used and in some
    tiu> cases, are outside the bounds of 0 and 1.

    tiu> I was wondering if anyone has considered in detail
    tiu> whether a "meta" estimator in which one takes the
    tiu> median amongst all estimators after excluding any which
    tiu> do not fall into the [0,1] range has any known
    tiu> statistical properties ? Any other recommendations also
    tiu> welcome.

    tiu> Thanks in advance, Tolga

And where is the small reproducible example code?

What does fracdiff give?

More information about the R-SIG-Finance mailing list