Systematic evaluation of half-lives of proton emission, <i>α</i> decay, cluster radioactivity and spontaneous fission based on a Bayesian neural network

Abstract Systematic evaluation of the half-lives of four decay modes (proton emission, α decay, cluster radioactivity and spontaneous fission) toward a unified description are performed through the Bayesian neural network (BNN) method. In BNN training, the existing experimental half-lives of nuclei are adopted directly as the target values. As a result, all the predicated half-lives of the four decay modes agree well with the experimental data, and the BNN approach is possibly useful for evaluations of half-lives of decay modes even if rare/insufficient experimental data are available. In addition, the angular momentum ℓ plays an important role in the prediction.