Constrain variables and results using nlinfit

Hello,

Basically I am fitting a result vector Y to a matrix X of 4 variables. Whatever the values of X my results vector Y should only take values between 0 and 1. How can I translate this in Matlab so that the fitting knows it should aim this range?

Example:

X= [1 5 8 2; 2 3 6 2; 8 5 4 1; 6 3 1 7]; Y= [0.2; 0.3; 0.7; 0.5];

modelfun = @(F,X) F(1).*X(:,1) + F(2).*X(:,2) + F(3).*X(:,3) + F(4).*X(:,4) + F(5).*X(:,3).*X(:,4) + F(6);

F_fitted = nlinfit(X,Y,modelfun,[1 10 1 1 1 1]);

-------------------------------

Thanks for your time!

[Pooja Lalan]

If I understand correctly, you would like the fitted parameters to result in a curve that takes values only between 0 and 1. Ideally you shouldn’t have to do that if you have good quality and quantity of data samples. The fitted curve should model the underlying ‘true’ curve. Constraining the output would not be mathematically sound. While there are ways of post-processing the fitted curve, like clipping the output at 0 and 1, there is no direct way of telling MATLAB to satisfy this constraint for all X while performing the curve fit.

Is there a particular reason why you would like to do that? Could you provide more information about your specific use case?