Machine Learning in Differential Diagnosis Lead Investigator: Nihar Patel Institution : UC Berkeley E-Mail : NiharPatel@berkeley.edu Proposal ID : 669 Proposal Description: This project aims to use machine learning methods in differential diagnosis of dementia. Our first goal is to select and develop meaningful patient attributes in the data set. Second, we will select, edit, and train different machine learning models on the previously curated data set. Third, we will then test and evaluate the model model produced by machine learning. Fourth, We will develop a pipeline to repeat this process until sufficient accuracy is demonstrated, after which we will test the results on a previously-untouched fraction of the data set. Fifth, pending the success of this process, we will develop a web-based tool for entering NACC test results on current patients and providing the output of this new data from the developed model.