From: John Conover <john@email.johncon.com>
Subject: Re: Chaos v. survival (was Re: General Equilibrium Model)
Date: 29 Dec 1998 02:00:52 GMT
Don Libby writes:
>
> Thanks for the references. For your dissertation consider establishing
> the formal correspondence between survival analysis and fractal
> analysis. I'm intrigued by the possibility that fractal analysis has
> something new to offer, but I suspect it may simply be a reinvented
> wheel.
>
Oh, I think that is quite true. For example, the run lengths of
industrial markets, equity values, market indices, GDP expansion,
(ie., recession and depressions,) crime rates, etc., all seem to
follow a 1 / sqrt (t) frequency distribution fairly accurately. The
distribution is invariant on sampling frequency, (ie., whether the
time interval is days, weeks, or months, the run length statistics
don't change,) implying a fractal type of scenario.
I'm not saying there is a fractal under every bed, but run lengths are
a useful concept. (Using an FFT do derive the same statistics is a
technically superior methodology, but the concept of complex
frequencies requires mathematical sophistication, where run lengths
are somewhat intuitive.)
John
BTW, I tend to like run length analysis for fractal time series-the
programming is simple and straight forward, its intuitive, and there
are "cheats" to anticipate what effect limited data set sizes would
have on the analysis. There are some industrial market run length
analysis for the electronics market place at:
http://www.johncon.com/john/correspondence/981014184454.18095.html
http://www.johncon.com/john/correspondence/981014210544.18525.html
http://www.johncon.com/john/correspondence/981014222823.18931.html
http://www.johncon.com/john/correspondence/981014233807.19309.html
as some examples.
--
John Conover, john@email.johncon.com, http://www.johncon.com/