CODEWITHSUNDEEP
matlabmaximize
user3547227
user3547227

Reputation:

Matlab maximize function

I have a function that gives me a likelihood which I want to maximize. The problem is that I need to give 4 parameters that I need to estimate. I have this function:

[likelihood, z0, z1, z2, z3]= myfun(g, g1, g2, x1, g3, x2, x3)

that in output gives likelihood how can I maximize it? I know all the g but I do not know the x and I should estimate them with maximization of the likelihood. Moreover I do not know the z as well.

I guess I should do fminsearch but I could not find documentation with parameter estimation like this.

Upvotes: 1

Views: 255

Answers (1)

Will
Will

Reputation: 1880

fminsearch seems like it would do just fine here - you just have to make inputs that abstract the search problem in the way that function expects. Here I think this means creating a function that:

  1. Takes the unknown parameters as a single array argument
  2. Returns a value to minimize, not maximize

Assuming that 0 is a sensible initial value for all x parameters, this would look something like this:

x = fminsearch(@(x) -myfun(g, g1, g2, x(1), g3, x(2), x(3)), [0,0,0]);
x1 = x(1);
x2 = x(2);
x3 = x(3);

To find out the z values and resulting likelihood from the search result returned, you can then simply put the result back into myfun:

[likelihood, z0, z1, z2, z3] = myfun(g, g1, g2, x1, g3, x2, x3)

Upvotes: 1

Related Questions