Approaches to Principal Component Analysis

Linear Data

MNIST

• PCA using eigenvectors of the data covariance matrix:

  1. Using Linear Data
  2. On MNIST Dataset

• PCA for Complete MNIST Dataset using gradient descent:

  1. without regularization
  2. with L1 regularization
  3. with L2 regularization

Using the full MNIST training and testing sets

Go here for MNIST in a easy-to-use CSV format.

---