Curve fit in python using scipy.optimize.curve_fit

Question

I am trying to curve fit my data with scipy.optimize.curve_fit. For some reason it doesn't like my equation. If I plot the equation using plausible numbers it looks right. So I trust my equation. If I try to fit to a simpler equation, it works, so I trust my code in general. Any advice as to why it doesn't work? The specific error return is "Residuals are not finite in the initial point." Changing p0 to anything reasonable is also unhelpful

import matplotlib.pyplot as plt
from scipy.optimize import curve_fit
import numpy as np




Y=np.array([0.00001,1.421828697,2.553970883,3.340262844,3.931459808,
4.640896164,5.084293887,5.379892368,5.557251457,5.705050698,
5.823290091,5.911969636,5.971089332,6.059768876,6.118888573,6.178008269,
6.266687814])


X=np.array([0,2500,5000,7500,10000,15000,20000,25000,
30000,35000,40000,45000,50000,55000,60000,65000,70000])

plt.plot(X,Y,'g^')



def func(X,J,g):

    u=6.720*10**-5
    k=1.380*(10**-16)
    T=2
    N=1
    x=g*J*u*(X)/(k*T)
    return N*g*J*(((2*J+1)/(2*J))*(1/(np.tanh((2*J+1)*x/(2*J))))-(1/(2*J))*
    (1/(np.tanh(x/(2*J)))))




plt.figure(1)
popt, pcov = curve_fit(func, X,Y,p0=[3.5,2],bounds=([0.25,0.001],[10,4]))
plt.plot(X, func(X, *popt), 'r--', label='fit-with-bounds')

print(" J is %s " %(popt[0]),"\n","g is %s" %(popt[1]))

Is it referring to the bounds? Im not clear what the initial point is — notAI
– notAI, Commented Dec 8, 2017 at 1:50
What about reading the docs? Initial point is p0 and is set to 1 if not given. — Julien
– Julien, Commented Dec 8, 2017 at 1:54

James Phillips · Accepted Answer · 2017-12-08 10:10:04Z

1

You include an X value of exactly zero, causing divide-by-zero errors. If I replace the value of zero with 0.01 your code works in my tests.

answered Dec 8, 2017 at 10:10

James Phillips

4,6673 gold badges16 silver badges11 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Curve fit in python using scipy.optimize.curve_fit

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related