An introduction to Bootstrap for nuclear physics

TL;DR

This paper introduces bootstrap methods for error estimation and parameter inference in nuclear physics, demonstrating their advantages over traditional techniques through simple examples and correlation analysis.

Contribution

It provides a practical introduction to bootstrap techniques in nuclear physics, including error quantification and parameter estimation, with comparisons to standard likelihood methods.

Findings

Bootstrap effectively quantifies error bars in nuclear physics estimators.

Bootstrap offers improved confidence interval evaluation when accounting for correlations.

Comparison shows bootstrap's advantages over traditional likelihood-based methods.

Abstract

This guide aims at providing a general introduction to bootstrap methods. By using simple examples taken from nuclear physics, I discuss how such a method can be used to quantify error bars of an estimator. I also investigate the use of bootstrap to estimate parameters of a simple liquid drop model. In particular, I investigate how the method compares with standard techniques based on likelihood estimator and how to take into account correlations to better evaluate confidence interval of parameters.