Comparing IR System Components Using Beanplots

摘要

In this poster we demonstrate an approach to gain a better understanding of the interactions between search tasks, test collections and components and configurations of retrieval systems by testing a large set of experiment configurations against standard ad-hoc test collections.