HP Forums
41C - Curve Fitting Revisited - Printable Version

+- HP Forums (https://archived.hpcalc.org/museumforum)
+-- Forum: HP Museum Forums (https://archived.hpcalc.org/museumforum/forum-1.html)
+--- Forum: Old HP Forum Archives (https://archived.hpcalc.org/museumforum/forum-2.html)
+--- Thread: 41C - Curve Fitting Revisited (/thread-165435.html)



41C - Curve Fitting Revisited - Ángel Martin - 04-06-2010

I'm compiling some curve fitting programs into a dedicated module, and was wondering if there's any RAW file (or other loadable format) for William Kolb's work.

I've finally succeeded porting the AECROM program & MCODE function -quite a fast implementation on the NUT world - so I thought it'd be nice to complement it with the other Classic reference work on the same subject...

Best Easter wishes,
ÁM




Re: 41C - Curve Fitting Revisited - Geir Isene - 04-06-2010

This is of great interest to me.

I have pretty much given up on the AECrom due to strange conflicts with other roms I use (not label or XROM conflicts, but some weird buffer issues locking up the calc). So, I took Kolb's book and added some needed features like Momentums, Skewness, Kurtosis and basic statistics. I then produced the following FOCAL program:

STAT v. 0.2


Re: 41C - Curve Fitting Revisited - Ángel Martin - 04-07-2010

Quite a long one! :) Thanks Geir, I'd say I should send you the transplanted CURVE and CRVF functions and you try them out.

The AECROM uses the "available" bits in the c(13) register and that conflicts with other custom ROMS, like the CCD - which uses it for the WSIZE among other things. This may explain your troubles...

So, you'll have mail...


Re: 41C - Curve Fitting Revisited - Geir Isene - 04-07-2010

Thanks for the explanation regarding the conflicting use of the c-register. It clears up this issue for me. I prefer the CCD over the AECrom - although I really needed the curve-fit program, hence the long FOCAL program. Having it as part of a non-conflicting MCODE module would be the best.

The FOCAL program covers several needs:

1. Curve fitting with a large range of curves
2. Basic statistics
3. Advanced statistics such as Momentums/Skewness/Kurtosis
4. The ability to use a data set from XM and analyze it through #1-#3 above (both single and paired data points).

If we could produce a new STAT module that covers the above needs and others (ANOVA etc.) - with saved data sets (XM, cards or mass storage) I would be in heaven.


Edited: 7 Apr 2010, 12:19 p.m.


Re: 41C - Curve Fitting Revisited - Ángel Martin - 04-07-2010

That kind of statistical heaven you dream of :) sure will require a reliable method to transfer the program to the rom - typing it is out of the question!

The AECROM funcions work FAST (again, for Coconut standards) but don't offer the advanced stuff you added. I like the XM option, very neat!

cheers


Re: 41C - Curve Fitting Revisited - Paul Dale - 04-07-2010

Is there any documentation to go with this large FOCAL programme?

- Pauli


Re: 41C - Curve Fitting Revisited - PeterP - 04-08-2010

Angel,

Assuming Geir has the Focal program installed on one of his 41s, maybe he could be enticed to export it to his NoVram via one of the available OS (I think Hepax would do that as well). One could then export the module from the NoVram to a PC at which point it would be available in electronic form.

There is a little article I wrote which shows how one can mix and match MCODE and Focal programs in a module. That way we could mesh his program with your AEC extraction (BTW - that's just awesome, Angel!)

Cheers

Peter