|
PureBytes Links
Trading Reference Links
|
I use a program called TableCurve 2D (moderately expensive). It's a fairly
complete "automatic" curve-fitting program. It can produce a frequency
spectrum for a time series and also find best-fit parameters for a large
number of built-in equations or for any user-specified equation.
The company which produces it has changed hands a couple of times since I
purchased it, but I suspect that a search on the Internet should turn it up.
Carroll Slemaker
> Does anyone know of a piece of software that will let you take say, 30
years
> of commodity data and derive one or two dominant cycles and amplitudes
> that 'fit' the data best? I am not looking for a 'system', just basic
> analytical software that can try many different time cycle combinations
> until it finds the best fits to approximate the data curve - then REPORT
(if
> available) the phase, amplitude and length of these cycles...
>
> I asked a friend of mine who said:
>
> "I remember downloading from some web site software that would extract the
> top N cycles from a time series and then superimpose that result over the
> original series. Can't recall where I found it. Try Google on the
> keywords..." He said it used a text data file and was Windows based.
>
> I tried a search and found nothing. I use TS 4.0 and Dial Data EOD data -
> would be nice if it were compatible with TS 4.0 OR even a stand alone
> Windows based or DOS version would be fine!
>
> I realize this is a form of optimization, (gag) but I wanted to speed up
> some new cycle grunt work I'm doing by hand...
>
> Thanks!
>
> Tom Cathey
|