FYS-STK-4155-Project2-Neural-Networks
In this project we present a practicum of classification and regression using feed forward neural networks, with comparison to traditional methods such as logistic regression. We implement gradient descent methods and neural network regression on a simple quadratic function, and logistic regression and neural network classification on the Wisconsin Breast Cancer Dataset. Our report is a pdf in the 'report' folder. Our codes are in the 'code' folder. The code for the optimization methods section is in 'gradient_descent.ipynb'. The code for the regression and classification using neural networks (both from scratch and in Tensorflow) is in 'neural_network.ipynb' - this notebook contains a Python class with our neural network. The code for logistic regression is in 'logistic_regression.ipynb'. And a test of ordinary least squares regression on our simple quadratic function (for comparison) is in 'OLS_scikit_learn.ipynb'.