Data Fitting with Least Squares

In [2]: 
import numpy as np
import numpy.linalg as npla
import scipy.linalg as spla
import matplotlib.pyplot as pt
In [3]: 
a = 4
b = 2

def f(x):
    return a + b*x

plot_grid = np.linspace(-3, 3, 100)

pt.plot(plot_grid, f(plot_grid))
Out[3]:
[<matplotlib.lines.Line2D at 0x7f73f9d879e8>]
In [4]: 
npts = 5

np.random.seed(22)
points = np.linspace(-2, 2, npts) + np.random.randn(npts)
values = f(points) + 0.3*np.random.randn(npts)*f(points)

pt.plot(plot_grid, f(plot_grid))
pt.plot(points, values, "o")
Out[4]:
[<matplotlib.lines.Line2D at 0x7f73f9d47da0>]

What's the system of equations for $a$ and $b$?

Now build the system matrix $A$:

In [5]: 
A = np.array([
    1+0*points,
    points,
    ]).T
A
Out[5]:
array([[ 1.        , -2.09194992],
       [ 1.        , -2.46335065],
       [ 1.        ,  1.08179168],
       [ 1.        ,  0.76067483],
       [ 1.        ,  1.50887086]])

What's the right-hand side vector?

Now solve the least-squares system:

In [6]: 
Q, R = npla.qr(A, "complete")
In [7]: 
print(A.shape)
print(Q.shape)
print(R.shape)

m, n = A.shape

(5, 2)
(5, 5)
(5, 2)

Determine $x$. Use spla.solve_triangular(A, b, lower=False).

In [8]: 
x = spla.solve_triangular(R[:n], Q.T.dot(values)[:n], lower=False)

Recover the computed $a$, $b$:

In [9]: 
a_c, b_c = x
In [10]: 
def f_c(x):
    return a_c + b_c * x

pt.plot(plot_grid, f(plot_grid), label="true")
pt.plot(points, values, "o", label="data")
pt.plot(plot_grid, f_c(plot_grid), label="found")
pt.legend()
Out[10]:
<matplotlib.legend.Legend at 0x7f73f9c83198>
  • What should happen if we change the number of data points?
  • What happens if there are lots of outliers?
  • What should happen if we don't add noise?
  • What about a bigger model?
  • What about different functions in the model?