1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
|
"""Fit y=ax+b and compute error bands"""
from vedo import Text2D, DashedLine, show
from vedo.pyplot import plot, fit
import numpy as np
# np.random.seed(0)
# Generate some noisy data points along a line
a, b = (np.random.rand(2)-0.5)*10 # choose a and b
x = np.linspace(0, 15, 25)
y = a*x + b
noise = np.random.randn(len(x)) * 5 # create gaussian noise
# Plot the points and add the "true" line without noise
fig = plot(x, y+noise, '*k', title=__doc__, label='data')
fig += DashedLine(x, y).c('red5')
# Fit points and evaluate, with a bootstrap and Monte-Carlo technique,
# the correct error coeffs and error bands. Return a Line object:
pfit = fit(
[x, y+noise],
deg=1, # degree of the polynomial
niter=500, # nr. of MC iterations to compute error bands
nstd=2, # nr. of std deviations to display
)
fig += [pfit, pfit.error_band, *pfit.error_lines] # add these objects to fig
msg = f"Generated a, b : {np.array([a,b])}"\
f"\nFitted a, b : {pfit.coefficients}"\
f"\nerrors on a, b : {pfit.coefficient_errors}"\
f"\nave point spread: :sigma :approx {pfit.data_sigma:.3f} in y units"
msg = Text2D(msg, font='VictorMono', pos='bottom-left', c='red3')
fig.add_label("y = a:dotx+b", c='k', marker='-', mc='red4')
fig.add_label("ground truth", c='k', marker='--', mc='red5')
fig.add_label("1:sigma error band", c='k', marker='s', mc='grey')
fig.add_legend(pos=[1.07,1], vspace=2)
show(fig, msg, zoom=1.1).close()
|
