Singular Value Decomposition (svd) in SciPy - Mini Project: Build & Apply

Singular Value Decomposition (svd) with SciPy

📖 Scenario: Imagine you work at a movie streaming company. You have a small matrix showing how much each user liked certain movies. You want to break this matrix into simpler parts to understand patterns better.

🎯 Goal: You will use scipy.linalg.svd to break a matrix into three parts: U, S, and Vt. Then you will print these parts to see the result.

📋 What You'll Learn

Create a 3x3 matrix called ratings with exact values

Create a variable full_matrices to control the SVD output shape

Use scipy.linalg.svd with ratings and full_matrices

Print the matrices U, S, and Vt

💡 Why This Matters

🌍 Real World

SVD is used in recommendation systems, image compression, and noise reduction by breaking complex data into simpler parts.

💼 Career

Understanding SVD helps in data analysis roles, machine learning, and any job involving matrix factorization or dimensionality reduction.

1

Create the ratings matrix

Create a 3x3 matrix called ratings using a list of lists with these exact values: [[5, 3, 0], [4, 0, 0], [1, 1, 0]]

SciPy

# Create the ratings matrix
# Your code here

Need a hint?

Use a list of lists to create the matrix exactly as shown.

2

Set the full_matrices variable

Create a variable called full_matrices and set it to False to get reduced SVD matrices.

SciPy

ratings = [[5, 3, 0], [4, 0, 0], [1, 1, 0]]
# Set full_matrices to False
# Your code here

Need a hint?

This variable controls the shape of the output matrices from SVD.

3

Perform Singular Value Decomposition

Import scipy.linalg and use scipy.linalg.svd with ratings and full_matrices. Store the results in variables U, S, and Vt.

SciPy

ratings = [[5, 3, 0], [4, 0, 0], [1, 1, 0]]
full_matrices = False
# Import scipy.linalg and perform svd
# Your code here

Need a hint?

Use scipy.linalg.svd with the full_matrices argument.

4

Print the SVD matrices

Print the variables U, S, and Vt each on a separate line.

SciPy

import scipy.linalg

ratings = [[5, 3, 0], [4, 0, 0], [1, 1, 0]]
full_matrices = False

U, S, Vt = scipy.linalg.svd(ratings, full_matrices=full_matrices)
# Print U, S, and Vt
# Your code here

Need a hint?

Use three print statements, one for each variable.