References & Citations
Mathematics > Numerical Analysis
Title: First- and second-order error estimates in Monte Carlo integration
(Submitted on 17 Jul 2015 (v1), last revised 20 Jan 2016 (this version, v2))
Abstract: In Monte Carlo integration an accurate and reliable determination of the numerical intregration error is essential. We point out the need for an independent estimate of the error on this error, for which we present an unbiased estimator. In contrast to the usual (first-order) error estimator, this second-order estimator can be shown to be not necessarily positive in an actual Monte Carlo computation. We propose an alternative and indicate how this can be computed in linear time without risk of large rounding errors. In addition, we comment on the relatively very slow convergence of the second-order error estimate.
Submission history
From: Ronald Kleiss [view email][v1] Fri, 17 Jul 2015 16:50:57 GMT (85kb)
[v2] Wed, 20 Jan 2016 15:33:59 GMT (76kb)
Link back to: arXiv, form interface, contact.