Author 
Message 
bajnes@gmail.com science forum beginner
Joined: 18 Nov 2005
Posts: 1

Posted: Fri Nov 18, 2005 7:11 am Post subject:
Math. function maker urgently required!



Dear Dudes,
Urgently wanted:
Is any software , able to model neumerus X per (Y1, Y2, Y3,...Yn) date
values by a math. function.
e.g. : X= 7.75000E02 5.77500E01 1.07750E+00 1.77750E+00 2.07750E+00
Y1=1.4799E04 1.9206E05 1.1978E06 9.7149E10 3.4080E13
Y2=3.9492E04 6.0642E05 5.1197E06 1.6097E08 1.1566E10
Y3= ...
and so on.
Looking for a function (not polynomial, serie,... but a final algebraic
function like: Yn=a.Ln(b.X+c) or some final formula which can model
such behviors. No matter how many coefficient this function may have.
of course each Yn differs with other by some constants.
Already known that aimed eq. is not linear, and not sure if matlab
could do what we want, or?
Many thanks for your notice,
Looking forward to your reply,
Best regards,
Bernard Luis 

Back to top 


Robert B. Israel science forum Guru
Joined: 24 Mar 2005
Posts: 2151

Posted: Fri Nov 18, 2005 8:09 am Post subject:
Re: Math. function maker urgently required!



In article <1132297915.524340.236220@z14g2000cwz.googlegroups.com>,
<bajnes@gmail.com> wrote:
Quote:  Urgently wanted:
Is any software , able to model neumerus X per (Y1, Y2, Y3,...Yn) date
values by a math. function.
e.g. : X= 7.75000E02 5.77500E01 1.07750E+00 1.77750E+00 2.07750E+00
Y1=1.4799E04 1.9206E05 1.1978E06 9.7149E10 3.4080E13
Y2=3.9492E04 6.0642E05 5.1197E06 1.6097E08 1.1566E10
Y3= ...
and so on.
Looking for a function (not polynomial, serie,... but a final algebraic
function like: Yn=a.Ln(b.X+c) or some final formula which can model
such behviors. No matter how many coefficient this function may have.
of course each Yn differs with other by some constants.

Ln is not an algebraic function, a polynomial is. But anyway, you might
guess a form for your function with a couple of parameters and use linear or nonlinear
leastsquares, as appropriate, to fit it. Since the function values vary widely in
magnitude while apparently being positive, you might try to fit ln(Yn) rather than
Yn directly. Perhaps something like ln(Yn) = a X^2 + b X + c (which you can do by
linear leastsquares), resulting in Yn = exp(a X^2 + b X + c).
Quote:  Already known that aimed eq. is not linear, and not sure if matlab
could do what we want, or?

Matlab should be perfectly capable of this.
Robert Israel israel@math.ubc.ca
Department of Mathematics http://www.math.ubc.ca/~israel
University of British Columbia Vancouver, BC, Canada 

Back to top 


Huynh science forum beginner
Joined: 18 Nov 2005
Posts: 1

Posted: Fri Nov 18, 2005 8:57 am Post subject:
Re: Math. function maker urgently required!



you should use function: lsqcurvefit(...) of matlab. It is very useful. 

Back to top 


Google


Back to top 



The time now is Fri May 25, 2018 12:43 pm  All times are GMT

