Data Mining
This is a team project that uses good data grooming and machine learning methods to predict when a hospital patient might be readmitted. I demonstrate data cleaning, supervised learning, regularization, and model selection methods to make decent predictions. Click the image to learn more.
Statistical Modeling
This project uses various statistical techniques to estimate from data the generating probability distribution. I use maximum likelihood estimation, hypothesis testing, and Bayes factors among others to analyze data. Click the image to learn more.
Machine Learning Interpretability
This project focuses on developing and finding the limitations of interpretability techniques in machine learning. I show that representational similarity analysis, a statistical technique for comparing model representations, depends heavily on the measure of similarity it uses. Click the image to learn more.
Unsupervised Learning
This project uses unsupervised learning techniques to learn a generative model for crime statistics. I demonstrate data cleaning, Python coding, generative modeling methods, and graphic visualization techniques. Click the image to learn more.
Algorithm Analysis
This project examines why some types of prediction aggregation techniques perform better than others. I demonstrate facility with data analysis and generation, and probabilistic modeling through Markov Chain Monte Carlo schemes in Pytorch. Click the image to learn more.
