Many machine learning algorithms are beginning to incorporate tools from algebraic topology and differential geometry, which allow algorithms to work on small data samples, messy optimization functions, and other messiness found in real data. This talk/workshop focuses on two promising algorithms that improve penalized regression models–homotopy LASSO and DGLARS–and includes an R tutorial analyzing an open-source educational dataset. Bring your laptops (and your messy datasets) for a hands-on session!
Colleen Farrelly is a Data Scientist at Kaplan University/Purdue Global University in Ft. Lauderdale. She has shared the datasets for the talk here: https://www.researchgate.net/project/Miami-Data-Science-Meetup, in case you would like to familiarize yourself in advance.
Paid parking is available in the parking lot adjacent to the building.
