How to chose the initial values for the coefficeints in nlinfit

 How to chose the initial values for the coefficients in nonlinear fit:

say I am tying to fit using

       a=nlinfit(x,y,'modelfun',a)

with a is the initial values for the coefficients

say I have define vector a as a=[0.1 1 1 1 0.1]

how to assign these initial values and what are these matrix elements (0.1,1,1,1,0.1) indicates.?

Answer 1

[Robert Cumming]

The values of "a" known as beta in the functions help are the initial values that the model will use as parameters in your 'modelfun'.

The values are 100% dependent on your model - you can start with zeros, ones or values which are "likely" to be near the end result - it can effect the solution so think about it carefully in advance.