Using Machine Learning to Predict Surgical Site Infection Biron, Dustin creator Jain, Sukrit creator Stey, Paul creator Anand, Rajsavi creator Chen, Elizabeth S. creator Sarkar, Indra Neil creator text posters 2017 eng Biomedical Informatics All rights reserved Surgical site infection (SSI) is a rare, but serious complication for patients undergoing total joint replacements. We aimed to create a machine learning algorithm that could accurately diagnose infection in the 199 patients that had undergone total knee or hip replacements from the MIMIC-III database.1 After data preprocessing, processing, and analysis, we determined that the dataset was too small to produce a meaningful machine learning algorithm. However, we recognize that this study serves as an important framework for studying larger datasets with more patients. Machine learning Risk prediction