How to optimize a specific logistic type equation using 'fminunc' and quasi-newton algorithm ?

Hello, I have y (dependent variable) and x (independent variable) data and want to develop a logistic model in the form of y = ymax/(1+exp(b-k*x). I want to optimize the solution to get b and k value by writing a function and solving them using 'fminunc'.

The values are like: ydata = [1.5 1.8 2.1 2.5 2.65 2.75] with xdata = [1 3 5 7 10 20], ymax = max(ydata). I wrote following code: function f = glass(x,xdata, ydata) f = (2.06/(1+exp(x(1)-x(2)*xdata)))- ydata;

in command window: ydata = [1.5 1.8 2.1 2.5 2.65 2.75]; xdata = [1 3 5 7 10 20]; x0 = [0.5; 0.5]; F = @(x)glass(x,xdata, ydata); [x,fval] = fminunc(F,x0) but it fails every time and says too many input input argument. Also want to mention quasi-newton algorithm to optimize it. How could I solve this problem?

Any help is appreciating.

Answer 1

[Alan Weiss]

There is an example showing how to solve this type of problem with both fminunc and with other, more efficient solvers. Follow the example carefully (especially where it shows how to define the objective function).

Alan Weiss

MATLAB mathematical toolbox documentation