Grid-based probabilistic skyline retrieval on distributed uncertain data

Abstract

The skyline queries help users make intelligent decisions over complex data. It has been recently extended to the uncertain databases due to the existence of uncertainty in many real-world data. In this paper, we tackle the problem of probabilistic skyline retrieval on physically distributed uncertain data with low bandwidth consumption. The previous work incurs sharply increased communication cost when the underlying dataset is anti-correlated, which is the typical scenario that the skyline is useful. In this paper, we propose a knowledge sharing approach based on a novel grid-based data summary. By sharing the data summary that captures the global data distribution, each local site is able to identify large amounts of unqualified objects early. Extensive experiments on both efficiency and scalability have demonstrated that our approach outperforms the competitor.