fminunc: get Hessian approximation after each iteration

I am using fminunc with quasi-Newton algorithms (BFGS and DFP). Is there is a way to record the approximation of the Hessian after each iteration? I looked at using the Outputfcn field, but the optimValues structure doesn't seem to include the Hessian updates.

Answer 1

[Matt J]

You could try running fminunc 1 iteration at a time in a for-loop and saving the 6th output arguments,

  options.MaxIter=1;
  for i=1:N
     
   [x,~,~,~,~,hessian{i}] = fminunc(fun,x,options);

  end