This body of work is around systematically employing deep learning on a large number of classification and regression problems (mostly from the UC Irvine Machine Learning Repository) in the R programming language using the h2o machine learning package.
I'm in the process of building in automated optimization and model averaging, extending to regression problems, and defining various systematic analyses (e.g. new visualizations of networks) which would aim to turn into journal paper. Please feel free to fork and collaborate. Enjoy and please let me know if you have any questions/comments.
Due to the volume and size of datasets, I can't upload them to the repository. I had to do some pretty extensive cleaning and various modifications so if you would like to use the exact data please just send me an email.
Author John Mercer (analyticsurgeon at gmail.com)