cs181-homework5

Spring 2021 Machine Learning (CS 181) Homework 5

Problem Topics

Solutions contained in the personal-solutions folder

1. Expecation-Maximization for a Categorical-Geometric Mixture model
2. Apply Principal Components Analysis on image data and calculate the reconstruction error
3. Conditional independence properties of a Bayesian Network

---

Code

Implementation contained in the code folder

problem1-EM-Mixture-Model.py

• Referred to as T5_P1.py in the specifications
• Implements the Expectation-Maximization algorithm and tests the results on different sized samples from 3 components

problem2-Principal-Components-Analysis.py

• Referred to as T5_P2.py in the specifications
• Performs Principal Components Analysis on MNIST data