| Author |
 Topic  |
|
|
marasneo
Poland
9 Posts |
 Posted - 04/20/2015 : 11:08:42 AM
|
Hi!
I have tried to modify standard Origin ExpDecay2 function to solve equation like this:
Z=C0+C1*exp(x1*t)+C2*exp(x2*t)
where t is my x (independent variable) and Z is dependent (y)
I am interested in determination of x1 and x2 roots which should be always negative.
Unfortunately, after fitting, they are always positive. Probably I am doing something wrong...
There is my modified code below (probably with errors). I need to take into account also the offset x0 (I do not know if the formula will be the same as for ExpDecay2 or not).
Always before each fitting I am just setting the Auto Parameter Initialization and checking "fixed" for C0 and x0 offset values (the same for function like ExpDecay2 - is it ok?).
FUNCTION
C0 + C1*exp(-(x-x0)*ex1) + C2*exp(-(x-x0)*ex2)
PARAMETER INIT
x0 = min(x_data);
int sign;
ex1 = get_exponent(x_data, y_data, &C0, &C1, &sign);
if (ex1 > 0)
{
int nSize = x_data.GetSize();
C0 = y_data[0] + y_data[nSize-1] - C0;
C1 = C2 = -0.5 * sign * exp(C1+ex1*(x_data[0]+x_data[nSize-1])-x0*ex1);
ex1 = ex2 = 1 / ex1;
}
else
{
ex1 = ex2 = -1 / ex1;
C1 = C2 = 0.5 * sign * exp(C1 - x0/ex1);
}
Please, help me with this problem. |
|
|
JacquelineHe
287 Posts |
Posted - 04/21/2015 : 04:27:59 AM
|
Hi,
In Origin ExpDecay2 function, the parameter is /ex1.
But in you new function, that is *ex1.
So it can not follow the ExpDecay2 function initialization code of those parameters.
Thanks
Jacqueline
OriginLab
|
 |
|
| |
Topic |
|
|
|
Problem with my own exp fitting function