<![CDATA[The rapid advancement of data-driven technologies has increased the application of machine learning in predictive modeling, particularly in health-related domains. Accurate health score prediction requires effective handling of high-dimensional data, where irrelevant or redundant features may reduce model performance. This study proposes the use of the Bald Eagle Search (BES) algorithm for feature selection in health score prediction using a Random Forest regression model. The objective is to optimize the subset of input features in order to minimize prediction error while maintaining model stability. A quantitative experimental approach was employed by comparing a baseline Random Forest model with an optimized model using BES-based feature selection. Model performance was evaluated using Root Mean Square Error (RMSE) and the coefficient of determination (R²). The baseline model achieved an RMSE of 0.6113 and an R² of 0.9947, while the BES-optimized model obtained an RMSE of 2.4983 and an R² of 0.9126. Although the optimized model showed slightly lower performance, BES successfully reduced the feature space while maintaining competitive predictive capability. The results demonstrate the potential of metaheuristic-based feature selection in developing efficient and interpretable machine learning models for health score prediction.]]>
Copyrights © 2026
