RuntimeWarning: overflow encountered in exp Python/Scipy [duplicate]

Question

I'm a total beginner in using curve_fit() of scipy. I don't understand what the problem is in the following code of mine:

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

def func(x, a, b, c):
    return a * np.exp(-b * x) + c

xdata = [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]
ydata = [75,66,63,61,60,58,58,55,56,54,56,59,57,57,56,58,56,58,56,56,56]
popt, pcov = curve_fit(func, xdata, ydata)

It returns RuntimeWarning: overflow encountered in exp

Any idea what could be wrong? Thanks in advance!

Probably worth putting in a print statement into func to see what arguments it is being called with. — alani
– alani, Commented Aug 17, 2020 at 15:26

JohanC · Accepted Answer · 2020-08-17 17:00:43Z

As remarked in this post, np.exp quickly reaches overflow. You can avoid the overflow with adding bounds on b. Note that you only get a warning, and that the result of curve_fit isn't affected.

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

def func(x, a, b, c):
    return a * np.exp(-b * x) + c

xdata = [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]
ydata = [75, 66, 63, 61, 60, 58, 58, 55, 56, 54, 56, 59, 57, 57, 56, 58, 56, 58, 56, 56, 56]
popt, pcov = curve_fit(func, xdata, ydata, bounds=([-np.inf, 0.0001, -np.inf], [np.inf, 10, np.inf]))
xs = np.linspace(2, 22, 100)
plt.plot(xs, func(xs, *popt))
plt.scatter(xdata, ydata)
plt.show()

PS: Also note that the fitted function uses the datatype of x, which sometimes causes strange problems. In this example there isn't problem, but in general it could help to add xdata = np.array(xdata, dtype=float) or xdata = np.array(xdata, dtype=np.longdouble).

Collectives™ on Stack Overflow

RuntimeWarning: overflow encountered in exp Python/Scipy [duplicate]

1 Answer 1

Comments

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Linked

Related