First Place for Trakya University Students in Artificial Intelligence-Supported Healthcare Project
Trakya University students won first place with their patient fall risk prediction project at the competition held within the scope of the 3rd National Artificial Intelligence Meeting.
Students from Trakya University Keşan Yusuf Çapraz School of Applied Sciences achieved first place in the Undergraduate Students Poster Competition at the 3rd National Artificial Intelligence Meeting organized by Cappadocia University.
Yağmur Direk and Oğuzhan Bodur, students in the Department of Information Systems and Technologies at Trakya University, were deemed worthy of the award with their project titled "Artificial Intelligence-Supported Patient Fall Risk Prediction" prepared within the body of the Computer and Innovation Society. The project aims to increase patient safety by predicting potential patient falls in hospital inpatient wards in advance. The study analyzed patients' demographic information, mobility capabilities, laboratory results, and medication usage data.
In the system developed with XGBoost and Random Forest algorithms, the SHAP analysis method was utilized to make the decisions of the models explainable. The developed XGBoost model achieved an accuracy rate of 91.4 percent and a ROC-AUC score of 0.93, outperforming the traditional 76.2 percent Morse fall risk assessment scale.