# Non polynomial fit: how to proceed?

LPH shared this idea 8 years ago

Hi Murkle, Hi all,

Extending the conversation of the following thread, I haven’t found a satisfactory solution. My problem is very simple and banal in Physics: I want the student to recover... simply (without CAS) the adjustment values when a model is given.

Let me be more precise. Indeed, in the case of charging a capacitor or a free fall with friction (eg "the fall of Felix Baumgartner"), the phenomenon is modeled by an exponential fit! With GeoGebra , the phenomenon is modeled very easily with :

Fit[list1, a*(1- e^(-b*x))]

Where a and b are initialized sliders. That works very well. :D

The problem is that you can’t recover the "adjusted values of a and b "! :cry: :cry:

Indeed , the commands

Coefficients[ <Polynomial> ] and Coefficients[ <Conic> ]

No longer work...

I just think it hasn't been programmed: are you agree? :confused:

Would it be possible to improve this command? :confused: :confused:

Phil

https://ggbm.at/566921 1

hello

try

b=-f''(0)/f'(0)

a=f'(0)/b

or

g = TaylorPolynomial[f, 0, 2]

coef = Coefficients[g]

b = -2 (Element[coef, 1] / Element[coef, 2])

a = Element[coef, 2] / b

saludos 1

Hi Juan Vicente,

hello

try

b=-f''(0)/f'(0)

a=f'(0)/b

or

g = TaylorPolynomial[f, 0, 2]

coef = Coefficients[g]

b = -2 (Element[coef, 1] / Element[coef, 2])

a = Element[coef, 2] / b

saludos

Thank you for your answer, but something similar has already been proposed by Noël and in a pedagogical way this solution is not satisfactory to be given to a student. :( :(

I just want the student to use the GeoGebra Coefficients[] tools, already used for other approaches and test the new model. :anguished:

I wanted to know if it is very difficult to improve the Coefficients[] command (very simple in this case)? I think the challenge is to retrieve the values ​​of internal variables that are displayed, no?

Thanks again Juan Vicente for your so quick answer, even if it’s not completely satisfactory.

Phil 1

hello

OK, i think now i understand you

you need some like this for a general case

i hope developers do it

saludos

you can move the points

https://ggbm.at/566933 1

Hi Juan Vicente,

Thank you: that's a nice workaround. :D

hello

OK, i think now i understand you

you need some like this for a general case

i hope developers do it

I hope so. It would be greatfull! :confused:

Cheers,

Phil 1

Would it be possible to improve this command?

Yes, good idea! 1

Hi Murkle,

Thank you very much! It’s OK for me.

It’s a nice improvement for the students.

Thanks for them. :D

Cheers,

Phil