CMStatistics 2015: Start Registration
View Submission - CMStatistics
B0298
Title: Parallel and other simulations in R made easy: An end-to-end study Authors:  Marius Hofert - University of Waterloo (Canada) [presenting]
Martin Maechler - ETH Zurich (Switzerland)
Abstract: The world of copulas is highly non-linear. Many applications require simulation studies involving copulas. From a computational point of view, these studies can become quite demanding. In order to tackle such large-scale simulations, the Rpackage ``simsalapar'' has been developed. This package aims at simplifying statistical simulation studies and carefully deals with important task such as parallel computing, seeding, catching of warnings and errors, and measuring runtime. The approaches in ``simsalapar'' may be of interest to students, researchers and practitioners as a how-to for conducting realistic, large-scale simulation studies in R. A practical (copula) problem from the realm of Quantitative Risk Management serves as a motivating example.