24小时热门版块排行榜

返回列表

li_qun

铜虫 (著名写手)

应助: 2 (幼儿园)
金币: 61
散金: 759
红花: 3
帖子: 1240
在线: 419.7小时
虫号: 692320
注册: 2009-01-14

[求助] 为什么拟合结果总为所设初始值？

用这样一个函数：yy=1./(1+5.*b./(2.*k)+2.*b.^2./k.^2+b.^3./(2.*k.^3)).*b.*exp(b.*xx).*(xx<0)+(b./k+2.*b.^2./k.^2+b.^3./k.^3)./(4+10.*b./k+8.*b.^2./k.^2+2.*b.^3./k.^3).*k.*(k.*xx+2./(b./k+1)).^2.*exp(-k.*xx).*(xx>=0)去拟合（xx,yy）
myfittype=fittype('1./(1+5.*b./(2.*k)+2.*b.^2./k.^2+b.^3./(2.*k.^3)).*b.*exp(b.*xx).*(xx<0)+(b./k+2.*b.^2./k.^2+b.^3./k.^3)./(4+10.*b./k+8.*b.^2./k.^2+2.*b.^3./k.^3).*k.*(k.*xx+2./(b./k+1)).^2.*exp(-k.*xx).*(xx>=0)','dependent',{'yy'},'independent',{'xx'},'coefficients',{'b','k'});
myfit=fit(xx',yy',myfittype,'start',[4.2e9,1.4e9])
plot(myfit,xx,yy)

myfit =

   General model:
   myfit(xx) = 1./(1+5.*b./(2.*k)+2.*b.^2./k.^2+b.^3./(2.*k.^3)).*b.*exp(b.*xx)
                  .*(xx<0)+(b./k+2.*b.^2./k.^2+b.^3./k.^3)./(4+10.*b./k+
                  8.*b.^2./k.^2+2.*b.^3./k.^3).*k.*(k.*xx+2./(b./k+1)
                  ).^2.*exp(-k.*xx).*(xx>=0)
   Coefficients (with 95% confidence bounds):
   b =    4.2e+09  (3.72e+09, 4.68e+09)
   k =    1.4e+09  (1.369e+09, 1.431e+09)

拟合时设定了初始值，但无论初始值设为多少，拟合结果就是多少，请问问题出在那里了呢？

回复此楼