一段时间以来,我一直在尝试为某些数据拟合函数,scipy.optimize.curve_fit
但我遇到了真正的困难。我真的看不出这不起作用的任何原因。
# 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
这是由上面的代码生成的绘图图像:
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