Trees and forests in nuclear physics

Abstract

We present a simple introduction to the decision tree algorithm using some examples from nuclear physics. We show how to improve the accuracy of the classical liquid drop nuclear mass model by performing feature engineering with a decision tree. Finally, we apply the method to the Duflo–Zuker model showing that, despite their simplicity, decision trees are capable of improving the description of nuclear masses using a limited number of free parameters.

A New Mass Model for Nuclear Astrophysics: Crossing 200 keV Accuracy

“…A more detailed statistical test can be performed on these residuals to verify that they do not follow a regular Gaussian distribution (see, for example, Refs. [20,38] for more details) but for the current discussion, a qualitative analysis is sufficient.…”

Extrapolating from neural network models: a cautionary tale

“…input-layer variables) in order to maximise the quantity of information one can extract from the data. See discussion in reference [27] for more details.…”

“…The LD model expresses the nuclear binding energy, B, as a sum of five different terms depending uniquely on proton (Z) and neutron (N) number as where A = N + Z and the coefficients a v , a s , ... have been adjusted in [43]. features the model are simple one could them to directly train the [27]. here no aprioriknowledge of the model use only N and features.…”

A new mass model for nuclear astrophysics: crossing 200 keV accuracy

“…A more detailed statistical test can be performed on these residuals to verify that they do not follow a regular Gaussian distribution -see for example Refs. [15,33] for more details -but for the current discussion a qualitative analysis is sufficient.…”