Variance estimation in neutron coincidence counting using the bootstrap method

In the study, we demonstrate the implementation of the ”bootstrap” method for a reliable estimation of the statistical error in Neutron Multiplicity Counting (NMC) on plutonium samples. The ”bootstrap” method estimates the variance of a measurement through a re-sampling process, in which a large number of pseudo-samples are generated, from which the so called bootstrap distribution is generated. The outline of the present study is to give a full description of the bootstrapping procedure, and to validate, through experimental results, the reliability of the estimated variance. Results indicate both a very good correspondence between the measured variance and the variance obtained through the bootstrap method (average error of 6% and maximal error of 11% on the statistical variance), and a robustness of the method with respect to the duration of the mea- surement and the bootstrap parameters.