Re: Power law process

[Rod Price]

I tried to respond to this earlier, but my laptop (I was traveling)
wouldn't send outgoing email.

Try looking at a book called "Signal Processing with Fractals:
A Wavelet-Based Approach," by Gregory Wornell (Prentice-
Hall 1996).  The book is about 1/f processes, which are signals
with power-law statistics.  You seem to be looking for a single
random variable with power-law statistics, but you might be
able to shoe-horn the approach advocated in this book into
your application.

The process works by first generating a set of wavelet coefficients
randomly.  The variance of the random wavelet coefficients
changes with the wavelet scale in a manner related to the power
law you are looking for.  Doing the inverse wavelet transform
generates the signal.

If you have a canned wavelet transform around somewhere, this
is really very easy.

-Rod Price
TRW Denver

Sven N. Thommesen wrote:

> At 04:14 PM 11/7/1998 -0500, you wrote:
> >Hi Sven-
> >I haven't seen any such algorithms myself; there might be one in one of
> >these books:
>
>   <snip>
>
> >If all else fails, you can just implement Per Bak's sandpile simulation
> >and use the distribution generated by that. :)
> >-Ted
>
> Thanks, Ted,
>   I was looking for an easy answer; looks like I'll need to go pound sand ;-)
>
> -Sven
>
>                   ==================================
>    Swarm-Support is for discussion of the technical details of the day
>    to day usage of Swarm.  For list administration needs (esp.
>    [un]subscribing), please send a message to <address@hidden>
>    with "help" in the body of the message.




                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.