Sampling, Scoring, and Precision

Publication year: 2012
Source: Exploring Monte Carlo Methods, 2012, Pages 69-96

William L., Dunn , J. Kenneth, Shultis

 Summary: Methods for obtaining samples from both continuous and discrete PDFs are reviewed. These include the inverse CDF method, the rejection method, the composition method, the rectangular-wedge-tail decomposition method, the composition-rejection method, and the ratio of uniforms method. Next, features about scoring MC histories and tests for assessing the MC results are discussed. These test include the relative error, the figure of merit, variance of the variance, and estimation of the underlying PDF. The use of history weights is introduced to correct scores for biased MC games. The chapter ends with a discussion of accuracy and precision and the effects of…