Cutting Edge Technology
Artificial intelligence and machine learning are ubiquitous technologies in today’s healthcare field. With applications across a broad spectrum, from wearable devices and new diagnostic algorithms to identifying inefficiencies in healthcare administration, artificial intelligence is helping to change the way we deliver healthcare. But the technology can be difficult to understand, so Thomas E. McKee, Ph.D., professor in the College of Health Professions, is teaching students at MUSC how artificial intelligence and data analytics work to solve many of today’s problems.
“I believe that applying theory to real-world problems enhances a student’s education,” said McKee.
McKee has pioneered a hypothetical case study at the Norwegian School of Economics in Bergen, Norway. Students study the fictional company Sharp Edge, Inc. to compare traditional statistical techniques to artificial intelligence techniques. In the working scenario, Sharp Edge is being sued over alleged gender bias in promotions over the previous five years. Students assume the role of consultants to help the defense team. Using company data on all previous promotions, students attempt to develop an advanced analytics model which predicts past promotions without using gender as a variable.
Students use the R language, a popular coding language in machine learning, to develop and test models using decision trees, regression analysis and neural networks.
“Students’ understanding of the three analytic methodologies is greatly increased because they must load the data, set up the commands, run the commands, and interpret results for each of the models,” said McKee.“This brings them close to real-world data science.”
Following a critical evaluation of the data, the students learn that the neural network solution is 100% accurate at predicting promotions with gender data compared to the regression model, which is 92% accurate.
McKee plans to incorporate this case study into his classes at MUSC. Furthermore, McKee has several other case studies under development that use very large data sets to facilitate students finding interesting patterns, anomalies or other results.