Machine learning predictions for cross-sections of 43,44 Sc radioisotope production by alpha-induced reactions on Ca target

Serkan Akkoyun,Cafer Mert Yeşilkanat,Tuncay Bayram
DOI: https://doi.org/10.1016/j.nimb.2024.165293
2024-02-25
Abstract:43,44 Sc radioisotopes are an alternative to 18 F in positron emission tomography. 43,44 Sc radioisotopes, which can be generated at low costs by irradiating inexpensive natural Ca with alpha particles, can be produced, and distributed in a central cyclotron facility due to their relatively long half-lives. Since there is limited experimental data on the cross-sections in the literature, in this study, cross-section predictions of the production of 43,44 Sc radioisotopes with alpha particles on Ca target were carried out with different machine learning approaches. In order to improve the results, the feature engineering method was applied to the variables of the cross-section predictions. Moreover, predictions have been improved with Stacked Ensemble Learning (SEL) approaches, a complex methodology that leverages the predictive capabilities of multiple underlying models to build a higher-level metamodel. We found that the best results were obtained with Bayesian Regularized Neural Network, Support Vector Regression and Stacked Ensemble Learning methods.
physics, nuclear, atomic, molecular & chemical,nuclear science & technology,instruments & instrumentation
What problem does this paper attempt to address?