This assignment is released as a Python Notebook at this link. Please read the instructions there and complete the required parts.

In this assignment you will need to implement the Singular Value Decomposition (SVD) algorithm from scratch. You will also need to use the SVD to perform dimensionality reduction on a dataset, explore the effect of the number of dimensions on the performance of a classifier, and visualize your results for comparison and analysis.

You can use Google Colab to edit the Python notebook or run your local Jupyter server.

Once the Python notebook is completed, follow these steps to submit:

1. Create a private repo called bu_username-assignment-3 following the repo naming convention.
2. Add the completed assignment3.ipynb to the repo (it can be the only file).
3. Make the bu_username-assignment-3 repo public on the deadline.