Getting Prime Cuts from Skylines over Partially Ordered Domains.

摘要: Skyline queries have recently received a lot of attention due to their intuitive query formulation: users can state preferences with respect several attributes. Unlike numerical preferences, over discrete value domains do not show an inherent total order, but rely on partial orders as stated by the user. In such typically many object values are incomparable, increasing size skyline sets significantly, and making computation expensive. this paper we explore how enable interactive tasks like refinement or relevance feedback providing ‘prime cuts’. Prime cuts interesting subsets full Pareto skyline, which give good overview skyline. They be small, efficient compute, suitable for higher numbers predicates, representative. The key improved performance reduced result set sizes is relaxation semantics concept weak dominance. We argue that yields results it opens up use scalable processing algorithms. Assessing practical impact, our experiments approach leads lean outperforms computations two magnitude.