Curve fitting program?

[sikyon]

Does anyone know of a good program to fit a parabola to data with? I want to fit a parabola of A*X^2 +B =0, not the general second order fitting of A*X^2 + B*X +C =0

[Haruki-kun]

I use one called "Winplot", but I don't know where to get it. I got it from my teacher.

[Brickwall]

I'm very good at fitting to curves. You don't need a program. :smallwink:

Oh, parabolas. I should read more carefully. :smalltongue:

[Shishnarfne]

Does anyone know of a good program to fit a parabola to data with? I want to fit a parabola of A*X^2 +B =0, not the general second order fitting of A*X^2 + B*X +C =0

For something as simple as a parabola, Excel might work (be warned, I've only used it for lines...), or a graphing calculator (TI series, for example), though those won't give you much precision, and you may have to check your fit.

I've worked with both WinCurveFit (I believe it has a relatively small fee) and Graphical Analysis (which might cost a bit more), which will fit most any curve (if you tell it which one you want) and give a "goodness of fit" assessment as well as the tolerance on the parameters...

But both of those are probably more than you need for a parabola...

Another alternative: plot against the square root of x instead of x... then you just get a line, so that might be simpler. You'll need to adjust your constants slightly, but you can verify the relationship easily.

[sikyon]

For something as simple as a parabola, Excel might work (be warned, I've only used it for lines...), or a graphing calculator (TI series, for example), though those won't give you much precision, and you may have to check your fit.

Excel won't do it



Another alternative: plot against the square root of x instead of x... then you just get a line, so that might be simpler. You'll need to adjust your constants slightly, but you can verify the relationship easily.

Mmm I'm not sure that the least-squares will extend back from linear to quadratic. Maybe they will, but if I recall from biochemistry this is a dangerous method to use, and is much less accurate especially due to weightings and stuff like that.

It's alright though, one of my co-workers has mathmatica and a proram to do it with. Thanks.

[Ponce]

If you're just plotting, then Graphmatica should work. Use Matlab if you are handy with matrix labs (or Octave if you are handy with matrix labs and are too cheap to pay for software, like me).