SVD with Scikit Learn

We can apply the SVD decomposition in Scikit Learn. Let’s see how we can get the U the Sigma and the V matrices.

from sklearn.decomposition import TruncatedSVD
import numpy as np

np.random.seed(0)
X = np.random.rand(100, 100)

# four components
svd = TruncatedSVD(n_components=4, n_iter=10, random_state=5)

U = svd.fit_transform(X)
Sigma = np.diag(svd.singular_values_)
V = svd.components_

In case we want to do the matrix multiplication, we can do it as follows:

# in case we want to do the multiplication

# U x Sigma
U_x_Sigma = np.dot(U, Sigma)
 
# (U x Sigma) x V
U_Sigma_V = np.dot(U_x_Sigma , V)

You can have a look at the SVD with SciPy.