A Bayesian nonparametric procedure for comparing algorithms

A fundamental task in machine learning is to compare the performance of multiple algorithms. This is typically performed by frequentist tests (usually the Friedman test followed by a series of multipl