Question: Nonlinear regression

Hello everybody,

I have an nx2 matrix of (t,w) , I need to fit a curve of the form of w = A(t-b)c to the data. What I have already tried is to pick a value for b, then regress log w on log (t-b) linearly, which is not very accurate. I'd  greatly appreciate any comments that could help me to find the optimum values for A, b and c through regression methods.

Thanks a lot,

Haley

Please Wait...