Classification and Regression, from linear and logistic regression to neural networks

This repository contains programs, test runs, material and report for project 2 in FYS-STK4155 made in collaboration between Oda Hovet (odasho), Ilse Kuperus (ilseku) and Erik Alexander Sandvik (erikasan).

Structure

Report folder

• Contains the PDF of the report

Programs folder

• mylearn package with logistic regression and linear regression with optimization methods for gradient descent
• Code for neural network: neural_network.py
• Code for classification with neural network: classification_with_neural_network.py
• Notebook for gradient descent regression: GD_Regression.ipynb
• Code for loading MNIST data set: mnist_loader.py

Test runs folder

• The mylearn package to run the other programs.
• Tests for logistic regression: LogisticRegression.ipynb
• Tests for verifying gradient descent optimization for linear regression: verify_GDRegressor.ipynb
• Test runs for regression with neural network: regression_with_neural_network.py

Figures folder

• Contains figures used in the report