Prediction of neutron-induced fission product yields by a straightforward k -nearest-neighbor algorithm

Abstract

Machine learning as a very powerful tool has recently been applied in many nuclear aspects. In this paper, a straightforward $k$-nearest-neighbor algorithm (KNN) combined with the general description of the fission observable (gef) method is proposed as an easy machine learning algorithm to learn existing fission product yields (FPYs) and predict unknowns. The results show that the KNN $+$ gef method can properly take account of the similarities of fission processes and the systematic properties of FPYs. Due to KNN $+$ gef, to use the experimental FPY data from the several nearest-neighboring isotopes, this method can provide an effective tool to predict the FPY data of the nuclei for which the less experimental data are available.