Linear + Quadratic Discriminant Analysis (From Scratch)

• Only using NumPy, I developed a Quadratic Discriminant Analysis (QDA) classifier which identified handwritten digits with 82% accuracy

Description

• Only using NumPy and Sklearn’s train_test_split, I implemented a Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA) classifier from scratch

• Training on 50,000 handwritten digits from the MNIST dataset, and evaluating on a test set of 10,000 handwritten digits, I achieved 82% accuracy