Predictive polynomial modeling and validation of mass attenuation coefficients for SrO-doped bioglasses

Abstract Strontium Oxide (SrO)-doped bioglasses are essential in biomedical applications due to their enhanced bioactivity, biocompatibility, and radiation shielding properties. The Mass Attenuation Coefficient (MAC) plays a crucial role in understanding their radiation interaction, which is vital for medical imaging and therapy. Traditional methods for calculating MAC are often time-consuming and computationally intensive. In contrast, polynomial regression provides a computationally simpler and more time-efficient alternative. This study investigates the effect of partially substituting Phosphorus Pentoxide (P 2 O 5 ) with varying amounts of Strontium Oxide (SrO) in 63S37C bioglasses and examines their radiation shielding capabilities. Using both experimental data and XCOM database references, six bioglass compositions were analyzed across photon energies from 15 to 15,000 keV. Monte Carlo simulations using the GEANT4 toolkit were also employed to validate the predicted MAC values. Polynomial regression models are employed to predict MAC, capturing the nonlinear relationship between energy and MAC. The selection of the optimal degree of the polynomial model is critical. Error evaluation metrics such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared (R 2 ) were used to assess model performance and determine the most suitable degree of the polynomial. Additionally, an early stopping algorithm was applied to reduce computational requirements by preventing overfitting and ensuring efficient model training. The results indicate that the optimal polynomial degree varies between 7, 9, and 10 for different compositions, demonstrating the importance of selecting the appropriate degree to balance accuracy and computational efficiency. The findings highlight the efficiency of polynomial regression as a practical method for predicting MAC, offering advantages over traditional approaches that are resource-intensive. This approach contributes to more effective design and evaluation of bioglasses in biomedical and radiation applications.