Project 2- FYS-STK4155-Applied Machine Learning an data analysis

Description

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

Source code structure

• src/ffnn.py: Class containing the feed forward neural networks
• src/logistic_reg.py: Class containing the logistic regression functions
• src/utils.py: Utillity functions like activation functions and such
• src/logeg_predict.py: Performing logistic regression on the Wisconsin breast cancer dataset. To run the functions, scroll down to the bottom of the script and remove the hashtag in front of function to run.
• src/franke.py: Script containing Frank's function. Is used to produce the dataset in the regression case.
• src/sgd_predict.py: Performing regression on Franke's dataset by using the SGD method in OLS and Ridge.
• src/franke_results/franke_predict.py: Performing regression on Franke´s dataset by using the neural network.
• src/mnist_results/mnist_predict.py: Performing classification on the MNIST dataset by using the neural network.

Important report files

• doc/main.tex: The report as a tex file
• doc/references.bib: References as a bib file
• doc/main.pdf: The report as a pdf file
• doc/assets: A folder with all the figures

Authors:

• Ida Due-Sørensen
• Philip Karim Niane
• Knut Magnus Aasrud