wrong interpolation in scipy.interpolate.interp1d

Question

I have problem with scipy.interpolate.interp1d

df = 
ang frc
-40 -15605.47
-30 -16051.253
-20 -10895.4
-10 -6205.733
0   -1730.65
10  2697.85
20  7217.59
30  12022.48
40  17298.647

import pandas as pd
from scipy.interpolate import interp1d

df = pd.read_csv('111.txt', sep='\t')
print(df)
interpi = interp1d(df['frc'], df['ang'], fill_value='extrapolate')
print(interpi(-17000))

It gives import pandas as pd from scipy.interpolate import interp1d

df = pd.read_csv('111.txt', sep='\t') print(df) interpi = interp1d(df['frc'], df['ang'], fill_value='extrapolate') print(interpi(-17000))

It gives -8.717 which is clearly wrong!
I rather it gives Error

can anyone help?

Answer 1

I think this depends on your definition of "wrong". To view this problem from the perspective of the interpolator (or rather extrapolator): If x_new = - 17000 , ie, outside of the range, the following points (blue in the picture below) are used to extrapolate: (-15605.47, -40) and (-16051.253, -30) . From those, the interpolator learns that a decrease in frc leads to an increase in ang . See the following picture. I clipped interpolation range for visibility.