Machine Learning Driven Insights into Nuclear Shape Transitions And Collectivity In The Transitional 150 Mass Region, A LightGBM Approach

This study, for the first time, investigates shape transitions and nuclear collectivity in the 150 mass region using a machine learning (ML) approach based on the Light Gradient Boosting Machine (LightGBM) algorithm. By examining 35 isotopes and validating predictions with newly measured data, we address gaps in electromagnetic transition probability, a key framework for understanding nuclear shape transitions and collectivity, which are impacted by data limitations for certain isotopes. The ML model outperforms traditional nuclear models, such as the IBA-1 and density functional theory, in predicting nuclear properties, highlighting the importance of features like the Casten factor (P). Our findings demonstrate the potential of ML to decode complex nuclear correlations and provide deeper insights into quantum shape transitions and collectivity in heavy nuclei. This work demonstrates the future potential of ML in driving nuclear research, enabling accurate property predictions, guiding experimental designs, and exploring regions beyond current experimental reach.