머신러닝을 이용한 노인의 잔존치아 수 예측모형 개발: 고령화연구 패널조사 9차 자료를 활용하여

Development of Remaining Teeth Prediction Model for the Elderly Using Machine Learning: Using the 9th Data of the Aging Research Panel Survey

Objectives: This study aimed to develop a machine-learning model to predict the number of remaining teeth in the elderly and identify the most influential factors. Methods: Using the 9th wave of the Korean Longitudinal Study of Aging (KLoSA), 4,451 participants aged 65 and older were analyzed. Predictive models using XGBoost, Random Forest, and Logistic Regression were built based on demographic and health behavior variables. Results: XGBoost demonstrated the best performance (AUROC: 0.686, accuracy: 0.673, precision: 0.690, recall: 0.872, F1 score: 0.770). The top predictors were age, subjective health status, educational level, economic activity, and diabetes. Conclusions: The model has potential utility in identifying high-risk elderly populations and informing preventive oral health policies.