Introduction

Implementation of different optimization / classification algorithms. This was done as a research project for the MSc. in Computer Engineering. The different classifiers implemented are tested on two sets of data:

• MNIST: A set of 70K (vectorized) 28x28 pixel images depicting hand-written images
• ORL: A set of 400 (vectorized) 40x30 facial images depicting 40 persons

Algorithms implemented

The following classification algorithms are implemented:

• Nearest class centroid
• Nearest sub-class centroid
• Nearest neighbour
• Perceptron trained using backpropagation
• Perceptron trained using MSE

External libraries & requirements

This project uses the following embedded libraries:

• Eigen: A C++ template library for linear algebra: matrices, vectors, numerical solvers, and related algorithms.
• mnist-parser: A simple C++ reader for MNIST dataset