On the Estimation of Additive Models: A Comparison of Splines and Neural Networks

This research project is a simulation study. Many methods exist to estimate additive models. I compare two of these methods: splines and neural networks. A brief application to real data is included in the appendix.

​

Do Neural Networks Predict a Country’s Obesity Prevalence More Accurately than Additive Models?

This project proposes a predictive model of a country's obesity prevalence. I compare the predictive performance of additive models and feedforward, dense neural networks.

​

How Many Personality Types Are There?

In this project, I apply unsupervised machine learning techniques to personality test data. In particular, I test for "groups" of data points using clustering algorithms. I model these groups as personality types. Note that some PDF viewers do not fully display figure 2.

​