[FIXED] Strange result of numpy polyfit

Issue

I've got strange (from my point of view) result from numpy polyfit. My code:

import numpy as np
data=np.array([2482.9, 2483.0, 2485.9, 2486.0, 2486.4, 2485.1, 2485.4, 2484.9, 2484.8, 2484.8, 2484.8, 2484.0, 2484.1, 2484.1, 2484.1])
wr = range(len(data))
poly  = np.polyfit(wr , data, deg = 2)
wp  = np.poly1d(poly)
el = 2484.1
res = wp(el)
print(res)
#result -225256.888955

Is this a bug?

Solution

As @DSM has already said - it doesn't look like a quadratic polynom.

We can try to fit it with a higher degree though:

import numpy.polynomial.polynomial as poly
x = wr; y = data

coefs = poly.polyfit(x, y, 4)
ffit = poly.Polynomial(coefs)

plt.plot(x, y)
plt.plot(x, ffit(x))
plt.legend(['y(x)','ffit(x)'])

Result:

Answered By - MaxU - stand with Ukraine