[交流] 【求助】数据拟合

function F=myfun(a,data)
x(1)=data(1,:);
x(2)=data(2,:);
F=a(3)*exp((-a(5))*(x(2)-a(4))/(a(6)+(x(2)-a(4))))/(1+(a(3)*exp((-a(5))*(x(2)-a(4))/(a(6)+(x(2)-a(4))))*x(1)/a(1))^(1-a(2)))
命令窗口程序如下：
>>  x1=[40 60 80 100 150 200 300 400 500 600 800 1000 1200 40 60 80 100 150 200 300 400 500 600 800 1000 1200 40 60 80 100 150 200 300 400 500 600 800 1000 1200 40 60 80 100 150 200 300 400 500 600 800 1000 1200 40 60 80 100 150 200 300 400 500 600 800 1000 1200];
x2=[423.15,423.15,423.15,425.15,425.15,425.15,425.15,425.15,425.15,425.15,425.15,425.15,425.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,431.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,437.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,443.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15,449.15];
ydata=[500 367.4 316.4 351 272 249 207 180 128 121 86.2 80.8 69.1 172.9 142.4 122.1 114 95 84 69.2 60 53 48 41.2 36.2 31.6 152.9 122.4 107.1 98 84 71 58.2 51 45.6 41.8 36.4 32.7 30.0 132.5 115.4 101.9 90 76 69 57.1 51 44.9 40.8 35.2 31.0 28.2 112.1 101.9 86.7 82 68 61 50.3 45 41 38 32.6 29 26.2];
data=[x1;x2];
a0=[5000 0.3 3.5*10e13 300 20 1.5];
[a,resnorm]=lsqcurvefit(@myfun,a0,data,ydata)

Originally posted by wgy318 at 2010-09-14 09:09:29:
能不能给你一个例子或者一个大体的程序框架啊？matlab里面应该有优化函数的程序可以调用吧。我现在函数已经有了，只是想优化下拟合的程度。但是我matlab又不怎么懂，不知道该怎么进一步优化。请多多指点

function F=myfun(a,data)
xx=data(1,:);
xxx=data(2,:);
F=a(3).*exp((-a(5))*(xxx-a(4))./(a(6)+(xxx-a(4))))./(1+(a(3).*exp((-a(5)).*(xxx-a(4))./(a(6)+(xxx-a(4)))).*xx/a(1)).^(1-a(2)))