[Scilab-users] horner test

[Rafael Guerra]

(note: votre message n'a pas été publié)

Bonsoir,

Essayez le code suivant svp:

// START OF CODE
clear;
clf;

M = [
0.003       170000
0.0028      208980
0.0026      261261
0.0024      336000
0.0022      439530
0.002       570000
0.0018      785568
0.0016      1110375
0.0014      1667952
0.0012      2646000
0.001       4500000
0.0008      8812500
0.0006      20875000
0.0004      70500000 ];

x = M(:,1);
y = M(:,2);

function f = fmodel(a,x)
     f = exp(a(1) + a(2)*x + a(3)*x.^a(4));
endfunction

function f = costfun(a,x,y)
    f = norm(y - fmodel(a,x));
endfunction

a0 = [0;0;0;0];
[f,aop]= optim(list(NDcost,costfun,x,y),a0);
str = sprintf("Scilab''s optim result:  y = %.4f*exp(%.2f*x
%+.2f*x^%+.2f)\n",exp(aop(1)),aop(2),aop(3),aop(4));
plot(x,y,"*r",x, fmodel(aop,x),"blue")
h = gca();
h.title.text= str;
h.title.font_size = 4;
xgrid
printf("\nEquation to model data:  y = a1*exp(a2*x + a3*x^a4)\n")
printf("%s\n",str)
printf("Cost function value = %g",f)

// NOTE: the cost function value can be further reduced by fixing
a(4)=-0.021;
//       and inverting for only 3 parameters: a(1), a(2) and a(3)
// END OF CODE

Cordialement,
Rafael




--
View this message in context: http://mailinglists.scilab.org/horner-test-tp4034828p4034875.html
Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com.