An Ensemble Machine Learning Model for Osteoporosis Risk Prediction from Medical Data

The metabolic condition osteoporosis afflicted a large number of people worldwide. Osteoporosis causes major health and its financial implications worldwide. However, the advancement of osteoporosis-related methods for forecasting is inadequate. Machine Learning (ML), a subfield of Artificial Intelligence (AI)I allows computers to ‘learn’ using programs. In comparison to classic statistical methods, ML places a greater focus on prediction accuracy and can find patterns in multidimensional data sets. This study propose an ensemble model employing Random Forest(RF), XGBoost(XGB) and Light Gradient Boosting(LGBM) techniques for the prediction of osteoporosis from the medical data of an individual. The accuracy, precision, recall, and f1-score were computed to compare the model with four alternative machine learning models: Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN). The findings show that the ensemble model outperforms with an accuracy, precision, recall and F1-score of 93.5%, 92.8%, 94.2% and 93.5% respectively.