Calculating the gradient of a function

Hello. I want to calculate the gradient of this function at the point xc:

function MSE=mseFunction(alpha,beta,y,yS)
MSE = [alpha beta; y yS];
end
xc = [100; 102];
y =  20;
yS = 50;

how I should proceed. Thanks!

Answer 1

[Marco Morganti]

Hi Amine,

you could use gradient() along with symbolic variables to find the gradient of your function MSE().

syms parameters;
f = mseFunction(parameters);
g = gradient(f);

at this point you can evaluate g() at the desired point:

g_xc = eval(subs(g,xc));

I hope this helps

[amine&&]

Hello Marco, When I type the following code

syms x1 x2;
f = mseFunction(x1,x2,y,yS);

I get :

The following error occurred converting from sym to double:
Error using symengine (line 59)
DOUBLE cannot convert the input expression into a double array.
If the input expression contains a symbolic variable, use VPA.

Error in mseFunction (line 33)
    LES(i)=alpha.*MoyenneMobileCentreeS(i)+(1-alpha).*LES(i-1);

Moreover I inform you that the code I already use is just a simplification. In fact y and yS are column vectors.

[Walter Roberson]

You should never eval() a symbolic expression. Symbolic expressions are in a language that is slightly different than MATLAB. You can matlabFunction the result of gradient() and pass c to that.

[amine&&]

Hi Walter, even typing :

f = @(x)mseFunction(x(1),x(2),y,yS)
g = gradient(f)

I get :

Undefined operator '.'' for input arguments of type 'function_handle'.
Error in gradient>parse_inputs (line 146)
        f = f.';
Error in gradient (line 48)
[f,ndim,loc,rflag] = parse_inputs(f,varargin);

[Walter Roberson]

syms parameters;
f = mseFunction(parameters);
g = gradient(f);

gfun = matlabFunction(g);      %rather than eval()
g_xc = gfun(xc);