Co-chairs: Alastair Young and Gil González-Rodríguez
Description: This WG will collaborate with WG C for the development of Task 4. Cross-validation and bootstrapping are two of the most frequently employed resampling-based approaches in statistics and data analysis. They are based on the idea that available samples can be used to obtain other versions of possible samples from the underlying population, and frequently provide more accurate results at a higher computational cost. Their robust versions are less extended and even more challenging. Computational strategies based on the re-utilization of the previous operations can significantly reduce the computational burden. However, the combinatorial nature of the problems will make the novel approaches computationally arduous and even infeasible for large-scale and high-dimensional problems. Consistency results providing the methods with a sound basis will be proved.