Bayesian Inference Based on Gaussian Process for Studying the α-Decay Half-Life of Atomic Nuclei

Key Points

• The aim is to develop a new analytic framework that quantifies parameter uncertainties in the alpha decay half-life using Bayesian inference and Gaussian processes.
• Developed a new decay law formula integrating Bayesian inference and Gaussian processes.
• Conducted sensitivity and feature importance analyses to identify key parameters affecting decay predictions.
• Evaluated model performance using root mean square error (RMSE) on recent alpha decay data.
• Achieved an overall RMSE of 0.375, with improved accuracy for even-even nuclei at 0.186.
• Identified barrier penetration factor and nuclear mass number as the most significant contributors to decay predictions.
• Model generalized well with recent observations, enhancing parameter uncertainty quantification.