Machine Learning Model Predicts Osteoporosis Risk Using Chronic Disease Data
-
Aims to develop a machine learning model to predict osteoporosis risk using chronic disease data from a German national database.
-
Employs several ML algorithms and finds stacked ensemble of Logistic Regression, AdaBoost, and Gradient Boosting performs best.
-
Model achieves AUC of 0.76 on test data, demonstrating good predictive performance.
-
Identifies age, gender, lipid disorders, cancer, and COPD as top 5 risk factors for osteoporosis.
-
Proposes model could assist in early screening and detection of high osteoporosis risk individuals in community settings.