1.你这个数据显然不能用y=exp(-k*X^N)去拟合
当x=0时y=exp(-k*X^N)=1,而你给的第一个y值是64.66,这错的太离谱了!!!
2.用y=a*exp(-b*x^c)是可以的
General model:
f(x) = a*exp(-b*x^c)
Coefficients (with 95% confidence bounds):
a = 66.01 (63.28, 68.75)
b = 0.01589 (0.006259, 0.02552)
c = 0.8514 (0.7225, 0.9803)
Goodness of fit:
SSE: 22.64
R-square: 0.9887
Adjusted R-square: 0.9868
RMSE: 1.374
能否将
y=exp(-k*X^N)
改为
y=a*exp(-b*x^c)
你这不能用ployfit拟合,那是拟合线性曲线的,要用lsqcurvefit,lsqcurvefit是拟合指数函数的!
polyfit