Bootstrap resampling as a tool for calculating uncertainty measurement

Abstract

The objective of this paper is to provide an applied research comparing the traditional and bootstrap methods to calculate the measure uncertainty. For this purpose, were performed a dimension analysis for internal and external diameter on one lot with around one hundred parts from a Brazilian company. Following, were performed resamples with replacementbootstrap samplesfor each dimension obtained and then the uncertainty calculation. After that, it was concluded that the proposed method is the most appropriate, because decreases the bias of estimation when it works with small sample size, which is common on metrology works. The companies researched assert that they do not perform the uncertainty calculation on dimension