Renaming the variabel while Symbolic to function handle conversion with matlabFunction

I do have the symbolic equation which is given as

A5, A1,(xdata and ydata) are the only variables and rest of them are constants.

Objective: To convert symbolic into function handle with replacing scalar variable (A1, A5) as a Vector A (A(1), A(2)).

When I use matlabFunction(F1),

F1d=vpa(subs(F1,[k m_canti B L Gap rho mu omega],[k_val m_canti_val B_val L_val Gap_val GWM_density(1) xdata ydata]),6)

func = matlabFunction(F1)

I am getting the result as below

func = 
    @(A1,A5,xdata,ydata)ydata.*-1.0-1.0./sqrt(A5.*3.987559999999988e-1+A1.*sqrt((xdata.*9.968899999999994e+2)./ydata).*2.26274169979699e-4+1.073920000000006e-3).*4.260164316079795

Since I am going to use this function for curve fitting I want A1, A5 to be renamed as A(1) and A(2). Basically I want to introduce a vector A in the above function handle.

I have gone through the documentation of matlabFunction(), Still I couldn't able to understand how to do what I want to achieve. I would be so grateful, if this problem got solved.

Thank you.

Answer 1

[Star Strider]

To get the ‘A’ values as a vector, put them in square brackets [].

func = matlabFunction(F1d, 'Vars',{[A1,A5],xdata,ydata})

The result should be:

func = @(in1,xdata,ydata) REST_OF_THE_FUNCTION_CODE

The ‘in1’ argument is a row vector where:

in1(1,:) = A1

in1(2,:) = A5

.

[Saravanakumar Dharmaraj]

Can I substitute xdata and ydata in the func and convert into function of one variable? like

func=@(in1) REST_OF_THE_FUCTION_CODE

[Star Strider]

You could, however that might make things more difficult.

I would just leave the arguments the way they are.

So for example to use it with fminunc, the call would simply be:

x0 = [1, 2]; % Use Appropriate Values For The Initial Parameter Estimates

A = fminunc(@(a)func(a,xdata,ydata), x0) % Estimate Parameters

although it might be necessary to use norm with it:

A = fminunc(@(a)norm(func(a,xdata,ydata)), x0) % Estimate Parameters

Make appropriate changes to use it with other optimisation functions.

Answer 2

[James Tursa]

You could keep your existing code and create another function handle:

funcv = @(A,xdata,ydata)func(A(1),A(2),xdata,ydata)

[Saravanakumar Dharmaraj]

Dear @James Tursa, I am doing that currently.

Different models yield different final equations. In this case, I need to write the function handle multiple times (I don't know how many are required to come to some conclusion). So I am looking to automate the process.

[James Tursa]

Please show the code you are currently using (or a small example that exibits this behavior) for creating the function handle. You want some parts of it to be vectors that are currently individual variables? What parts of this building process do you want "automated"? Etc.

[Saravanakumar Dharmaraj]

I have updated my question. Please do have a look.

Currently A1, A5 are individual variables. Yes.

Please neglect the world "Automation". I want to find a script to convert individual variables into a vector in function handle.