[MATLAB混合编程] 数据拟合初值选取

[sdsanlengcao]

怎样选择a,b的初值，才能使结果收敛的更好

function f = curvefun1( x , tdata)
f =( x(1)*tdata.^(3/2) + x(2)*tdata );
end


tdata = load('700-1-depthSI.txt');
cdata=load('700-1-loadSI.txt');
for a=1000000:1000000:10000000
for b=-10:1:0
x0=[a,b];
x = lsqcurvefit(' curvefun1 ',x0,tdata,cdata)
a
b

end
end


%原始数据tdata =

5.3307e-007
7.1173e-007
8.9054e-007
5.2882e-007
5.2753e-007
8.8644e-007
8.8518e-007
5.2330e-007
5.2197e-007
1.2414e-006
1.4200e-006
1.2385e-006
1.4172e-006
1.2356e-006
1.2344e-006
8.7271e-007
1.0515e-006
1.2302e-006
1.9497e-006
1.9480e-006
1.9468e-006
1.7653e-006
1.7639e-006
2.3030e-006
1.9413e-006
1.7598e-006
1.9386e-006
……
cdata =

5.2342e-006
5.9606e-006
6.6329e-006
7.3346e-006
7.9906e-006
8.6807e-006
9.3248e-006
1.0082e-005
1.0751e-005
1.1356e-005
1.2108e-005
1.2801e-005
1.3490e-005
1.4204e-005
1.4853e-005
1.5522e-005
1.6207e-005
1.6902e-005
1.7520e-005
1.8317e-005
1.8922e-005
1.9614e-005
2.0285e-005
2.0981e-005
2.1675e-005
2.2346e-005
2.3020e-005
2.3727e-005
……