一段时间以来我一直在尝试使用函数来拟合某些数据scipy.optimize.curve_fit
但我确实有困难。我真的看不出这行不通的任何原因。
# encoding: utf-8
from __future__ import (print_function,
division,
unicode_literals,
absolute_import,
with_statement)
import numpy as np
from scipy.optimize import curve_fit
import matplotlib.pyplot as mpl
x, y, e_y = np.loadtxt('data.txt', unpack=True)
def f(x, a, k):
return (1/(np.sqrt(1 + a*((k-x)**2))))
popt, pcov = curve_fit(f, x, y, maxfev = 100000000)
mpl.plot(x, f(x, *popt), 'r-', label='Fit')
mpl.plot(x, y, 'rx', label='Original')
mpl.legend(loc='best')
mpl.savefig('curve.pdf')
print(popt)
# correct values which should be calculated
# a=0.003097
# k=35.4
Here is the plot-image which is produced by upper code:
![enter image description here](https://i.stack.imgur.com/94242.png)
data.txt:
#x y e_y
4.4 0.79 0.13
19.7 4.9 0.8
23.5 7.3 1.2
29.7 17 2.79
30.7 21.5 3.52
34 81 13.28
34.6 145 23.77
35.4 610 100
36.3 115 18.85
38.1 38 6.23
43.7 14 2.3
56.2 6.2 1.02
64.7 4.6 0.75
79.9 3.2 0.52
210 0.98 0.16