Neural Networks for Classification

May 2019

---

This repository accompanies the paper Neural Networks for Classification and contains the implementations of the code described therein.

---

Contents of this repository

• The file logistic_regression.ipynb contains the step-by-step development and application of a logistic regression model to a binary classification problem. A couple of interesting data generation algorithms are introduced as a by-product.
• The net/ directory contains the code defining the extendable neural network model, the majority of which can be found in layers.py and network.py. The file utils.py contains utility functions which can be used to assist with the preparation of data and training of models. The remaining files contain applications of various models to assorted problems.
• The data/ directory is used to store NumPy-format data files used in the KMNIST examples - see the separate README file in that directory for more information.
• The report/ directory contains the final paper in PDF format, as well as the TeX source code, accompanying figures (as generated by the code in net/), and project specification.