| T O P I C R E V I E W |
| vanhufau |
Posted - 08/02/2001 : 07:43:43 AM
I am trying to fit data with a Hillīs function. I have defined this function, and I canīt find any problem in the expression.
I have 12 parameters, but each time I try to fit, it says :"try to set parameter ... as fixed", and it works only when more than 7 parameters are fixed.
Why?????
My function is :
y=2*mu1*(x^a1-(x*(1.35275-0.36352*x)^2)^(-a1*b1))/(a1*x) + 2*mu2*(x^a2-(x*(1.35275-0.36352*x)^2)^(-a2*b2))/(a2*x) + 2*mu3*(x^a3-(x*(1.35275-0.36352*x)^2)^(-a3*b3))/(a3*x) + 2*mu4*(x^a4-(x*(1.35275-0.36352*x)^2)^(-a4*b4))/(a4*x)
|
| 2 L A T E S T R E P L I E S (Newest First) |
| rtoomey |
Posted - 08/07/2001 : 5:06:55 PM
Faustine,
An empty chi-sqr value might indicate that the initial parameter guesses were inadequate and have caused the fit to wander off (i.e. it isn't converging). Fitting with the Levenburg-Marquardt algorithm (as you are in our fitter in most cases) requires fairly reasonable initial parameter estimates.
In a situation like this, it would be best if you sent the data to Technical Support along with your initial parameter estimates. The function can then be scrutinized and suggestions can be made as to how to improve it and make the fit work.
- rtoomey
|
| vanhufau |
Posted - 08/06/2001 : 03:55:41 AM
The problem was with my initial values...
Thank you!
Now I have another problem : It doesnīt return me any value for the CHI^2...
?
Faustine |
