This repo shows work that I did with a research group at the University of Arizona during the course of my master's degree. It involves the use of structured medical data to create machine learning models that classify children's risk of an autism diagnosis, using various factors including physicians' notes, age, and prior diagnoses. The data itself is protected by a NDA but these files show the general approach I took to solve this problem. Please see the included project report for a more detailed explanation and for the results of the work.