Repository for code and other resourches for the JHU machine learning in economics reading group.
The goal of these notebooks is to provide an opportunity for hands-on learning of machine learning tools. Notebooks with "blank" at the end have sections removed which should be filled in.
Currently, these notebooks deal with simple neural nets coded up "by hand" using numpy. "OLSNN" introduces a simple neural net with one layer and one neuron, using a linear model as an example, as well as provides a refresher on gradient descent optimization. "multiple_neurons" uses nonlinear data and expands the width (number of neurons) in the hidden layer and adds some additional flexibility and optimization tools.