A pitfall in the use of extended likelihood for fitting fractions of pure samples in mixed samples
arXiv:0803.2711 · doi:10.1016/j.cpc.2008.10.008
Abstract
The paper elucidates, with an analytic example, a subtle mistake in the application of the extended likelihood method to the problem of determining the fractions of pure samples in a mixed sample from the shape of the distribution of a random variable. This mistake, which affects two widely used software packages, leads to a misestimate of the errors.
10 pages with 1 figure, to appear in Computer Physics Communications. New version including major changes and additions, resulting from discussion with journal reviewers