Theoretical predictions on α -decay properties of some unknown neutron-deficient actinide nuclei by machine learning

Neutron-deficient actinide nuclei provide a valuable window to probe heavy nuclear systems with large proton-neutron ratios. In recent years, several new neutron-deficient Uranium and Neptunium isotopes have been observed via α-decay spectroscopy [Z. Y. Zhang et al., Phys. Rev. Lett. 122, 192503 (2019); L. Ma et al., Phys. Rev. Lett. 125, 032502 (2020); Z. Y. Zhang et al., Phys. Rev. Lett. 126, 152502 (2021)]. In spite of these achievements, some neutron-deficient key nuclei in this mass region are still unknown in experiments. Machine learning algorithms have been applied successfully in different branches of modern physics. It is interesting to explore their applicability in α-decay studies. In this work, we propose a new model to predict α-decay energies and half-lives within the framework based on a machine learning algorithm called the Gaussian process. We first calculate the α-decay properties of the new actinide nucleus \begin{document}$ {}^{214}{\rm{U}}$\end{document} . The theoretical results turn out to be in good agreement with the latest experimental data, which shows the reliability of our model. We continue to use the model to predict α-decay properties of some unknown neutron-deficient actinide isotopes and compare the results with traditional models. The results could be useful for future synthesis and identification of these unknown isotopes.