Matt Baucum is an assistant professor of business analytics in the Department of Business Analytics, Information Systems and Supply Chain at FSU's College of Business.
mbaucum@business.fsu.edu

A pioneering study from researcher Matthew Baucum, published in Management Science, demonstrates that, when paired with data-driven reinforcement learning, wearable sensor data yields insights that could help physicians better treat their patients.  

Sensor data might, for instance, suggest to a doctor that a patient's symptoms are most severe in the evening, Baucum said, and reinforcement learning might indicate that higher medication dosages in the afternoon can prevent severe evening symptoms. 

“The pairing of both can therefore inform a better treatment plan for the patient,” he says.

An assistant professor in the Department of Business Analytics, Information Systems and Supply Chain, Baucum continues to deepen his expertise in machine learning, healthcare analytics and insight on improved patient treatment. In another 2023 study, published in ACM Transactions on Management Information Systems, Baucum shows ways in which healthcare practitioners can use reinforcement learning – a form of artificial intelligence that uses data to learn optimal problem-solving strategies – to  personalize and optimize treatment for substance use disorder.

Baucum says his research “focuses on using big healthcare data sets to figure out the best way to treat patients and to allocate healthcare resources in a way that benefits as many people as possible.”

Recent publications or presentations:

Matt Baucum, Anahita Khojandi, Rama Vasudevan, Ritesh Ramdhani (2023) Optimizing Patient-Specific Medication Regimen Policies Using Wearable Sensors in Parkinson’s Disease. Management Science 69(10):5964-5982. https://doi.org/10.1287/mnsc.2023.4747

Matt Baucum, Anahita Khojandi, Carole Myers, and Larry Kessler. 2023. Optimizing Substance Use Treatment Selection Using Reinforcement Learning. ACM Trans. Manage. Inf. Syst. 14, 2, Article 13 (March 2023), 30 pages. https://doi.org/10.1145/3563778